23 Commits
Author SHA1 Message Date
Tanguy Pruvot 5fd8c32b4e some changes related to wait stats 2015-06-05 21:10:22 +02:00
Tanguy Pruvot 94a04aea96 multipool: fix some params like time-limit 2015-06-04 12:39:22 +02:00
Tanguy Pruvot 8ce8961e06 Increase max password size to 128 chars (nicehash)
--pass="f0=0;f2=0;f3=2.7;f4=1.9;f5=99;f6=1.8;f7=7.9;f8=0;f9=0.45;f10=0;f11=4.3;f12=3.9"
2015-06-04 12:39:01 +02:00
Tanguy Pruvot 34c04cfb80 multipool: prevent last longpoll http errors on switch 2015-05-26 10:19:02 +02:00
Tanguy Pruvot 6c101af7a2 multipool: allow failover switch from getwork 2015-05-26 09:50:31 +02:00
Tanguy Pruvot e606887e7a fix for vstudio 2015-05-25 15:08:38 +02:00
Tanguy Pruvot d4191d3188 stratum: check stratum_gen_work result
and remove useless purge test...
2015-05-25 07:36:29 +02:00
Tanguy Pruvot a20dc6a658 algo aliases, fix max-diff param without pool config 2015-05-25 03:16:09 +02:00
Tanguy Pruvot 69852a2874 stratum: properly free jobs on disconnect 2015-05-23 04:27:32 +02:00
Tanguy Pruvot ee9c821525 stratum: do not alloc empty merkle tree 2015-05-23 03:55:49 +02:00
Tanguy Pruvot 6fd935c369 multipool: remove last rpc_ vars in stratum thread 2015-05-22 11:24:04 +02:00
Tanguy Pruvot 2d6b3fddf6 multipool: increase stability, but not 100% perfect
several problems fixed:
- submit to the pool set in work (source pool)
- longpoll curl timeout could be too high and could lock the switch
- mutexes cant be copied on windows (stratum global var to fully remove later)

I linked the -T timeout option to curl and tried to remove all fixed timeout values
2015-05-22 08:32:08 +02:00
Tanguy Pruvot 0e6168ed52 Prepare the 1.6.4 release... 2015-05-22 08:31:59 +02:00
Tanguy Pruvot a9d3c1ffdb multipool: last changes before squashed merge
and fix net diff on wallets.. was longpoll specific
2015-05-21 07:19:50 +02:00
Tanguy Pruvot a63f0024f3 update api sample for the protocol 1.5 2015-05-21 05:28:09 +02:00
Tanguy Pruvot adda14b22e stats: store pool number in scanlog 2015-05-21 04:46:06 +02:00
Tanguy Pruvot e1a0274b01 api: switchpool and new pool stats variables (API v1.5)
add accepted/rej by pool, wait time on conditional, net diff and rate

also add scantime per pool config option and do some pool cleanup..
2015-05-21 04:05:22 +02:00
Tanguy Pruvot 1a30450ad2 prevent concurrent pool switching
and limit condtionnal wait messages to the first thread/device
2015-05-20 06:45:30 +02:00
Tanguy Pruvot e3922c7a33 add some pool rotate options, like pool time-limit
update sample pools.conf for time rotation
2015-05-20 05:58:04 +02:00
Tanguy Pruvot 312bd90541 do not try to restart threads from threads
Start inconditionally the stratum and longpoll threads,
these threads are just waiting a tq_push() if unused...

so add some checks to know if vars are set for the right pool
2015-05-20 04:50:40 +02:00
Tanguy Pruvot d4a9428fef pools: add name and removed attributes
also increase max defined pools to 8

to be tested on windows..
2015-05-20 01:07:45 +02:00
Tanguy Pruvot d840d683ec parse json config file pools array 2015-05-19 23:17:37 +02:00
Tanguy Pruvot d6c29b1f7f multi-pools: prepare storage/switch of credentials
for the moment:
- allow the storage of multiple -o params (and user/pass)
- allow a failover pool on connection failed
- switch to the next pool with the "seturl" api remote command
- longpoll to stratum switch (reverse to check...)

todo: mix stratum/getwork, new api commands, json config...
2015-05-19 12:48:02 +02:00
68 changed files with 2946 additions and 11862 deletions

No files matched your search

+13 -10
View File
@@ -15,7 +15,7 @@ static __device__ uint64_t cuda_swab32ll(uint64_t x) {
return MAKE_ULONGLONG(cuda_swab32(_LODWORD(x)), cuda_swab32(_HIDWORD(x)));
}
__constant__ static uint32_t c_data[3+1];
__constant__ static uint32_t c_data[20];
__constant__ static uint32_t sigma[16][16];
static uint32_t c_sigma[16][16] = {
@@ -195,16 +195,17 @@ void blake256_gpu_hash_80(const uint32_t threads, const uint32_t startNonce, uin
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
const uint32_t nonce = startNonce + thread;
uint32_t h[8];
uint32_t input[4];
#pragma unroll
for (int i = 0; i < 8; i++) h[i] = cpu_h[i];
#pragma unroll 8
for (int i = 0; i<8; i++) { h[i] = cpu_h[i];}
#pragma unroll
for (int i = 0; i < 3; ++i) input[i] = c_data[i];
#pragma unroll 3
for (int i = 0; i < 3; ++i) input[i] = c_data[16 + i];
input[3] = startNonce + thread;
input[3] = nonce;
blake256_compress2nd(h, input, 640);
#pragma unroll
@@ -229,14 +230,16 @@ void blake256_cpu_hash_80(const int thr_id, const uint32_t threads, const uint32
__host__
void blake256_cpu_setBlock_80(uint32_t *pdata)
{
uint32_t h[8], data[20];
uint32_t h[8];
uint32_t data[20];
memcpy(data, pdata, 80);
memcpy(h, c_IV256, sizeof(c_IV256));
for (int i = 0; i<8; i++) {
h[i] = c_IV256[i];
}
blake256_compress1st(h, pdata, 512);
cudaMemcpyToSymbol(cpu_h, h, sizeof(h), 0, cudaMemcpyHostToDevice);
cudaMemcpyToSymbol(c_data, &data[16], sizeof(c_data), 0, cudaMemcpyHostToDevice);
cudaMemcpyToSymbol(c_data, data, sizeof(data), 0, cudaMemcpyHostToDevice);
}
__host__
+116 -99
View File
@@ -20,17 +20,21 @@ static __constant__ uint2 vSKEIN_IV512_256[8] = {
{ 0x33EDFC13, 0x3EEDBA18 }
};
static __constant__ int8_t ROT256[8][4] = {
46, 36, 19, 37,
33, 27, 14, 42,
17, 49, 36, 39,
44, 9, 54, 56,
39, 30, 34, 24,
13, 50, 10, 17,
25, 29, 39, 43,
8, 35, 56, 22,
static __constant__ int ROT256[8][4] =
{
46,36, 19, 37,
33,27, 14, 42,
17,49, 36, 39,
44, 9, 54, 56,
39,30, 34, 24,
13,50, 10, 17,
25,29, 39, 43,
8, 35, 56, 22,
};
static __constant__ uint2 skein_ks_parity = { 0xA9FC1A22,0x1BD11BDA};
static __constant__ uint64_t skein_ks_parity64 = 0x1BD11BDAA9FC1A22ull;
static __constant__ uint2 t12[6] = {
{ 0x20, 0 },
{ 0, 0xf0000000 },
@@ -40,19 +44,25 @@ static __constant__ uint2 t12[6] = {
{ 0x08, 0xff000000 }
};
static __constant__ uint2 skein_ks_parity = { 0xA9FC1A22, 0x1BD11BDA };
static __constant__ uint64_t skein_ks_parity64 = 0x1BD11BDAA9FC1A22ull;
static __constant__ uint64_t t12_30[6] = {
0x20,
0xf000000000000000,
0xf000000000000020,
0x08,
0xff00000000000000,
0xff00000000000008
};
static __forceinline__ __device__
void Round512v35(uint2 &p0, uint2 &p1, uint2 &p2, uint2 &p3, uint2 &p4, uint2 &p5, uint2 &p6, uint2 &p7, int ROT)
{
p0 += p1; p1 = ROL2(p1, ROT256[ROT][0]) ^ p0;
p2 += p3; p3 = ROL2(p3, ROT256[ROT][1]) ^ p2;
p4 += p5; p5 = ROL2(p5, ROT256[ROT][2]) ^ p4;
p6 += p7; p7 = ROL2(p7, ROT256[ROT][3]) ^ p6;
p0 += p1; p1 = ROL2(p1, ROT256[ROT][0]); p1 ^= p0;
p2 += p3; p3 = ROL2(p3, ROT256[ROT][1]); p3 ^= p2;
p4 += p5; p5 = ROL2(p5, ROT256[ROT][2]); p5 ^= p4;
p6 += p7; p7 = ROL2(p7, ROT256[ROT][3]); p7 ^= p6;
}
static __forceinline__ __device__
void Round_8_512v35(uint2 *ks, uint2 *ts,
uint2 &p0, uint2 &p1, uint2 &p2, uint2 &p3,
@@ -62,36 +72,39 @@ void Round_8_512v35(uint2 *ks, uint2 *ts,
Round512v35(p2, p1, p4, p7, p6, p5, p0, p3, 1);
Round512v35(p4, p1, p6, p3, p0, p5, p2, p7, 2);
Round512v35(p6, p1, p0, p7, p2, p5, p4, p3, 3);
p0 += ks[(R+0) % 9]; /* inject the key schedule value */
p1 += ks[(R+1) % 9];
p2 += ks[(R+2) % 9];
p3 += ks[(R+3) % 9];
p4 += ks[(R+4) % 9];
p5 += ks[(R+5) % 9] + ts[(R+0) % 3];
p6 += ks[(R+6) % 9] + ts[(R+1) % 3];
p7 += ks[(R+7) % 9] + make_uint2((R),0);
p0 += ks[((R)+0) % 9]; /* inject the key schedule value */
p1 += ks[((R)+1) % 9];
p2 += ks[((R)+2) % 9];
p3 += ks[((R)+3) % 9];
p4 += ks[((R)+4) % 9];
p5 += ks[((R)+5) % 9] + ts[((R)+0) % 3];
p6 += ks[((R)+6) % 9] + ts[((R)+1) % 3];
p7 += ks[((R)+7) % 9] + make_uint2((R),0);
Round512v35(p0, p1, p2, p3, p4, p5, p6, p7, 4);
Round512v35(p2, p1, p4, p7, p6, p5, p0, p3, 5);
Round512v35(p4, p1, p6, p3, p0, p5, p2, p7, 6);
Round512v35(p6, p1, p0, p7, p2, p5, p4, p3, 7);
p0 += ks[(R+1) % 9]; /* inject the key schedule value */
p1 += ks[(R+2) % 9];
p2 += ks[(R+3) % 9];
p3 += ks[(R+4) % 9];
p4 += ks[(R+5) % 9];
p5 += ks[(R+6) % 9] + ts[(R+1) % 3];
p6 += ks[(R+7) % 9] + ts[(R+2) % 3];
p7 += ks[(R+8) % 9] + make_uint2(R+1, 0);
p0 += ks[((R)+1) % 9]; /* inject the key schedule value */
p1 += ks[((R)+2) % 9];
p2 += ks[((R)+3) % 9];
p3 += ks[((R)+4) % 9];
p4 += ks[((R)+5) % 9];
p5 += ks[((R)+6) % 9] + ts[((R)+1) % 3];
p6 += ks[((R)+7) % 9] + ts[((R)+2) % 3];
p7 += ks[((R)+8) % 9] + make_uint2((R)+1, 0);
}
__global__ __launch_bounds__(256,3)
void skein256_gpu_hash_32(uint32_t threads, uint32_t startNounce, uint64_t *outputHash)
{
const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint2 h[9], t[3] = { t12[0], t12[1], t12[2] };
uint2 h[9];
uint2 t[3];
uint2 dt0,dt1,dt2,dt3;
uint2 p0, p1, p2, p3, p4, p5, p6, p7;
h[8] = skein_ks_parity;
for (int i = 0; i<8; i++) {
@@ -99,23 +112,27 @@ void skein256_gpu_hash_32(uint32_t threads, uint32_t startNounce, uint64_t *outp
h[8] ^= h[i];
}
uint2 dt0 = vectorize(outputHash[thread]);
uint2 dt1 = vectorize(outputHash[threads + thread]);
uint2 dt2 = vectorize(outputHash[threads*2 + thread]);
uint2 dt3 = vectorize(outputHash[threads*3 + thread]);
t[0]=t12[0];
t[1]=t12[1];
t[2]=t12[2];
uint2 p0 = h[0] + dt0;
uint2 p1 = h[1] + dt1;
uint2 p2 = h[2] + dt2;
uint2 p3 = h[3] + dt3;
uint2 p4 = h[4];
uint2 p5 = h[5] + t[0];
uint2 p6 = h[6] + t[1];
uint2 p7 = h[7];
LOHI(dt0.x,dt0.y,outputHash[thread]);
LOHI(dt1.x,dt1.y,outputHash[threads+thread]);
LOHI(dt2.x,dt2.y,outputHash[2*threads+thread]);
LOHI(dt3.x,dt3.y,outputHash[3*threads+thread]);
#pragma unroll 9
p0 = h[0] + dt0;
p1 = h[1] + dt1;
p2 = h[2] + dt2;
p3 = h[3] + dt3;
p4 = h[4];
p5 = h[5] + t[0];
p6 = h[6] + t[1];
p7 = h[7];
#pragma unroll
for (int i = 1; i<19; i+=2) {
Round_8_512v35(h, t, p0, p1, p2, p3, p4, p5, p6, p7, i);
Round_8_512v35(h,t,p0,p1,p2,p3,p4,p5,p6,p7,i);
}
p0 ^= dt0;
@@ -141,19 +158,18 @@ void skein256_gpu_hash_32(uint32_t threads, uint32_t startNounce, uint64_t *outp
t[0] = t12[3];
t[1] = t12[4];
t[2] = t12[5];
p5 += t[0];
p5 += t[0]; //p5 already equal h[5]
p6 += t[1];
#pragma unroll 9
#pragma unroll
for (int i = 1; i<19; i+=2) {
Round_8_512v35(h, t, p0, p1, p2, p3, p4, p5, p6, p7, i);
}
outputHash[thread] = devectorize(p0);
outputHash[threads + thread] = devectorize(p1);
outputHash[threads*2 + thread] = devectorize(p2);
outputHash[threads*3 + thread] = devectorize(p3);
outputHash[thread] = devectorize(p0);
outputHash[threads+thread] = devectorize(p1);
outputHash[2*threads+thread] = devectorize(p2);
outputHash[3*threads+thread] = devectorize(p3);
}
}
@@ -161,10 +177,10 @@ static __forceinline__ __device__
void Round512v30(uint64_t &p0, uint64_t &p1, uint64_t &p2, uint64_t &p3,
uint64_t &p4, uint64_t &p5, uint64_t &p6, uint64_t &p7, int ROT)
{
p0 += p1; p1 = ROTL64(p1, ROT256[ROT][0]) ^ p0;
p2 += p3; p3 = ROTL64(p3, ROT256[ROT][1]) ^ p2;
p4 += p5; p5 = ROTL64(p5, ROT256[ROT][2]) ^ p4;
p6 += p7; p7 = ROTL64(p7, ROT256[ROT][3]) ^ p6;
p0 += p1; p1 = ROTL64(p1, ROT256[ROT][0]); p1 ^= p0;
p2 += p3; p3 = ROTL64(p3, ROT256[ROT][1]); p3 ^= p2;
p4 += p5; p5 = ROTL64(p5, ROT256[ROT][2]); p5 ^= p4;
p6 += p7; p7 = ROTL64(p7, ROT256[ROT][3]); p7 ^= p6;
}
static __forceinline__ __device__
@@ -175,36 +191,38 @@ void Round_8_512v30(uint64_t *ks, uint64_t *ts, uint64_t &p0, uint64_t &p1, uint
Round512v30(p2, p1, p4, p7, p6, p5, p0, p3, 1);
Round512v30(p4, p1, p6, p3, p0, p5, p2, p7, 2);
Round512v30(p6, p1, p0, p7, p2, p5, p4, p3, 3);
p0 += ks[(R+0) % 9]; /* inject the key schedule value */
p1 += ks[(R+1) % 9];
p2 += ks[(R+2) % 9];
p3 += ks[(R+3) % 9];
p4 += ks[(R+4) % 9];
p5 += ks[(R+5) % 9] + ts[(R+0) % 3];
p6 += ks[(R+6) % 9] + ts[(R+1) % 3];
p7 += ks[(R+7) % 9] + R;
p0 += ks[((R)+0) % 9]; /* inject the key schedule value */
p1 += ks[((R)+1) % 9];
p2 += ks[((R)+2) % 9];
p3 += ks[((R)+3) % 9];
p4 += ks[((R)+4) % 9];
p5 += ks[((R)+5) % 9] + ts[((R)+0) % 3];
p6 += ks[((R)+6) % 9] + ts[((R)+1) % 3];
p7 += ks[((R)+7) % 9] + R;
Round512v30(p0, p1, p2, p3, p4, p5, p6, p7, 4);
Round512v30(p2, p1, p4, p7, p6, p5, p0, p3, 5);
Round512v30(p4, p1, p6, p3, p0, p5, p2, p7, 6);
Round512v30(p6, p1, p0, p7, p2, p5, p4, p3, 7);
p0 += ks[(R+1) % 9]; /* inject the key schedule value */
p1 += ks[(R+2) % 9];
p2 += ks[(R+3) % 9];
p3 += ks[(R+4) % 9];
p4 += ks[(R+5) % 9];
p5 += ks[(R+6) % 9] + ts[(R+1) % 3];
p6 += ks[(R+7) % 9] + ts[(R+2) % 3];
p7 += ks[(R+8) % 9] + R+1;
p0 += ks[((R)+1) % 9]; /* inject the key schedule value */
p1 += ks[((R)+2) % 9];
p2 += ks[((R)+3) % 9];
p3 += ks[((R)+4) % 9];
p4 += ks[((R)+5) % 9];
p5 += ks[((R)+6) % 9] + ts[((R)+1) % 3];
p6 += ks[((R)+7) % 9] + ts[((R)+2) % 3];
p7 += ks[((R)+8) % 9] + (R)+1;
}
__global__ __launch_bounds__(256, 3)
void skein256_gpu_hash_32_v30(uint32_t threads, uint32_t startNounce, uint64_t *outputHash)
{
const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint64_t h[9], t[3];
uint64_t h[9];
uint64_t t[3];
uint64_t dt0, dt1, dt2, dt3;
uint64_t p0, p1, p2, p3, p4, p5, p6, p7;
h[8] = skein_ks_parity64;
for (int i = 0; i<8; i++) {
h[i] = SKEIN_IV512_256[i];
@@ -215,21 +233,21 @@ void skein256_gpu_hash_32_v30(uint32_t threads, uint32_t startNounce, uint64_t *
t[1] = devectorize(t12[1]);
t[2] = devectorize(t12[2]);
uint64_t dt0 = outputHash[thread];
uint64_t dt1 = outputHash[threads + thread];
uint64_t dt2 = outputHash[threads*2 + thread];
uint64_t dt3 = outputHash[threads*3 + thread];
dt0 = outputHash[thread];
dt1 = outputHash[threads+thread];
dt2 = outputHash[2*threads+thread];
dt3 = outputHash[3*threads+thread];
uint64_t p0 = h[0] + dt0;
uint64_t p1 = h[1] + dt1;
uint64_t p2 = h[2] + dt2;
uint64_t p3 = h[3] + dt3;
uint64_t p4 = h[4];
uint64_t p5 = h[5] + t[0];
uint64_t p6 = h[6] + t[1];
uint64_t p7 = h[7];
p0 = h[0] + dt0;
p1 = h[1] + dt1;
p2 = h[2] + dt2;
p3 = h[3] + dt3;
p4 = h[4];
p5 = h[5] + t[0];
p6 = h[6] + t[1];
p7 = h[7];
#pragma unroll 9
#pragma unroll
for (int i = 1; i<19; i += 2) {
Round_8_512v30(h, t, p0, p1, p2, p3, p4, p5, p6, p7, i);
}
@@ -254,22 +272,22 @@ void skein256_gpu_hash_32_v30(uint32_t threads, uint32_t startNounce, uint64_t *
h[8] ^= h[i];
}
t[0] = devectorize(t12[3]);
t[1] = devectorize(t12[4]);
t[2] = devectorize(t12[5]);
t[0] = t12_30[3];
t[1] = t12_30[4];
t[2] = t12_30[5];
p5 += t[0]; //p5 already equal h[5]
p6 += t[1];
#pragma unroll 9
#pragma unroll
for (int i = 1; i<19; i += 2) {
Round_8_512v30(h, t, p0, p1, p2, p3, p4, p5, p6, p7, i);
}
outputHash[thread] = p0;
outputHash[threads + thread] = p1;
outputHash[threads*2 + thread] = p2;
outputHash[threads*3 + thread] = p3;
outputHash[threads + thread] = p1;
outputHash[2 * threads + thread] = p2;
outputHash[3 * threads + thread] = p3;
} //thread
}
@@ -288,7 +306,6 @@ void skein256_cpu_hash_32(int thr_id, uint32_t threads, uint32_t startNounce, ui
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
// only 1kH/s perf change between kernels on a 960...
if (device_sm[device_map[thr_id]] > 300 && cuda_arch[device_map[thr_id]] > 300)
skein256_gpu_hash_32<<<grid, block>>>(threads, startNounce, d_outputHash);
else
+1 -32
View File
@@ -8,35 +8,4 @@ need to be installed on your system for autogen.sh to work)
where your CUDA 6.5 toolkit is installed (usually /usr/local/cuda,
but some distros may have a different default location)
** How to compile on Fedora 22 (2015) **
# Step 1: gcc and dependencies
dnf install gcc gcc-c++ automake
dnf install jansson-devel openssl-devel libcurl-devel zlib-devel pciutils-devel
# Step 2: nvidia drivers (Download common linux drivers from nvidia site)
dnf install kernel-devel
modprobe -r nouveau
./NVIDIA-Linux-x86_64-346.72.run
# Step 3: CUDA SDK (Download from nvidia archive)
dnf install cuda-repo-fedora20-6-5-prod-6.5-19.x86_64.rpm
dnf install cuda-minimal-build-6-5.x86_64
# add the nvcc binary path to the system
ln -s /usr/local/cuda-6.5 /usr/local/cuda
echo 'export PATH=$PATH:/usr/local/cuda/bin' > /etc/profile.d/cuda.sh
# add the cudart library path to the system
echo /usr/local/cuda/lib64 > /etc/ld.so.conf.d/cuda.conf
ldconfig
# Step 4: Fix the toolkit warning with gcc 5.1
edit the file /usr/local/cuda/include/host_config.h
and comment/delete the line 82 : #error -- unsupported GNU version! gcc 4.9
./build.sh
./ccminer -n
See README.txt
+4 -2
View File
@@ -2,7 +2,6 @@
#include <memory.h>
#include "cuda_helper.h"
#include "miner.h"
// ZR5
__constant__ uint32_t d_OriginalData[20];
@@ -10,6 +9,9 @@ __constant__ uint32_t d_OriginalData[20];
__constant__ uint32_t c_PaddedMessage[18];
__constant__ uint64_t c_State[25];
#define POK_DATA_MASK 0xFFFF0000
#define POK_VERSION 0x1
#define U32TO64_LE(p) \
(((uint64_t)(*p)) | (((uint64_t)(*(p + 1))) << 32))
@@ -655,7 +657,7 @@ __host__
void zr5_keccak512_cpu_hash_pok(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t* pdata, uint32_t *d_hash, uint16_t *d_poks)
{
const uint32_t threadsperblock = 256;
const uint32_t version = (pdata[0] & (~POK_DATA_MASK)) | (use_pok ? POK_BOOL_MASK : 0);
const uint32_t version = pdata[0] & (~POK_DATA_MASK);
dim3 grid((threads + threadsperblock-1)/threadsperblock);
dim3 block(threadsperblock);
+12 -13
View File
@@ -19,9 +19,8 @@ ccminer_SOURCES = elist.h miner.h compat.h \
compat/inttypes.h compat/stdbool.h compat/unistd.h \
compat/sys/time.h compat/getopt/getopt.h \
crc32.c hefty1.c \
ccminer.cpp pools.cpp util.cpp bignum.cpp \
ccminer.cpp util.cpp \
api.cpp hashlog.cpp nvml.cpp stats.cpp sysinfos.cpp cuda.cpp \
ethash/etherpc.cpp ethash/ethash.cu \
heavy/heavy.cu \
heavy/cuda_blake512.cu heavy/cuda_blake512.h \
heavy/cuda_combine.cu heavy/cuda_combine.h \
@@ -48,15 +47,15 @@ ccminer_SOURCES = elist.h miner.h compat.h \
sph/cubehash.c sph/echo.c sph/luffa.c sph/sha2.c sph/shavite.c sph/simd.c \
sph/hamsi.c sph/hamsi_helper.c sph/sph_hamsi.h \
sph/shabal.c sph/whirlpool.c sph/sha2big.c sph/haval.c \
qubit/qubit.cu qubit/qubit_luffa512.cu qubit/deep.cu qubit/luffa.cu \
pluck/pluck.cu pluck/cuda_pluck.cu \
qubit/qubit.cu qubit/qubit_luffa512.cu qubit/deep.cu qubit/doom.cu \
x11/x11.cu x11/fresh.cu x11/cuda_x11_luffa512.cu x11/cuda_x11_cubehash512.cu \
x11/cuda_x11_shavite512.cu x11/cuda_x11_simd512.cu x11/cuda_x11_echo.cu \
x11/cuda_x11_luffa512_Cubehash.cu \
x13/x13.cu x13/cuda_x13_hamsi512.cu x13/cuda_x13_fugue512.cu \
x15/x14.cu x15/x15.cu x15/cuda_x14_shabal512.cu x15/cuda_x15_whirlpool.cu \
x15/whirlpoolx.cu x15/cuda_whirlpoolx.cu \
x17/x17.cu x17/cuda_x17_haval512.cu x17/cuda_x17_sha512.cu \
x11/c11.cu x11/s3.cu
x11/s3.cu
# scrypt
ccminer_SOURCES += scrypt.cpp scrypt-jane.cpp \
@@ -75,8 +74,8 @@ ccminer_SOURCES += compat/winansi.c
endif
ccminer_LDFLAGS = $(PTHREAD_FLAGS) @CUDA_LDFLAGS@
ccminer_LDADD = @LIBCURL@ @JANSSON_LIBS@ @PTHREAD_LIBS@ @WS2_LIBS@ @CUDA_LIBS@ @PCILIB@ @OPENMP_CFLAGS@ @LIBS@ $(nvml_libs)
ccminer_CPPFLAGS = @LIBCURL_CPPFLAGS@ @PCIFLAGS@ @OPENMP_CFLAGS@ $(CPPFLAGS) $(PTHREAD_FLAGS) -fno-strict-aliasing $(JANSSON_INCLUDES) $(DEF_INCLUDES) $(nvml_defs)
ccminer_LDADD = @LIBCURL@ @JANSSON_LIBS@ @PTHREAD_LIBS@ @WS2_LIBS@ @CUDA_LIBS@ @OPENMP_CFLAGS@ @LIBS@ $(nvml_libs)
ccminer_CPPFLAGS = @LIBCURL_CPPFLAGS@ @OPENMP_CFLAGS@ $(CPPFLAGS) $(PTHREAD_FLAGS) -fno-strict-aliasing $(JANSSON_INCLUDES) $(DEF_INCLUDES) $(nvml_defs) -DSCRYPT_KECCAK512 -DSCRYPT_CHACHA -DSCRYPT_CHOOSE_COMPILETIME
nvcc_ARCH = -gencode=arch=compute_50,code=\"sm_50,compute_50\"
@@ -106,7 +105,7 @@ x11/cuda_x11_luffa512.o: x11/cuda_x11_luffa512.cu
$(NVCC) $(nvcc_FLAGS) --maxrregcount=80 -o $@ -c $<
x11/cuda_x11_luffa512_Cubehash.o: x11/cuda_x11_luffa512_Cubehash.cu
$(NVCC) $(nvcc_FLAGS) --maxrregcount=76 -o $@ -c $<
$(NVCC) $(nvcc_FLAGS) --maxrregcount=80 -o $@ -c $<
x13/cuda_x13_hamsi512.o: x13/cuda_x13_hamsi512.cu
$(NVCC) $(nvcc_FLAGS) --maxrregcount=72 -o $@ -c $<
@@ -128,21 +127,21 @@ JHA/cuda_jha_compactionTest.o: JHA/cuda_jha_compactionTest.cu
# This object does not use cuda device code but call the different kernels (autotune)
scrypt/salsa_kernel.o: scrypt/salsa_kernel.cu
$(NVCC) $(JANSSON_INCLUDES) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_21,compute_20\" -o $@ -c $<
$(NVCC) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_21,compute_20\" -o $@ -c $<
# These kernels are for older devices (SM)
scrypt/test_kernel.o: scrypt/test_kernel.cu
$(NVCC) $(JANSSON_INCLUDES) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_20,compute_20\" -o $@ -c $<
$(NVCC) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_20,compute_20\" -o $@ -c $<
scrypt/fermi_kernel.o: scrypt/fermi_kernel.cu
$(NVCC) $(JANSSON_INCLUDES) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_21,compute_20\" -o $@ -c $<
$(NVCC) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_20,code=\"sm_21,compute_20\" -o $@ -c $<
scrypt/kepler_kernel.o: scrypt/kepler_kernel.cu
$(NVCC) $(JANSSON_INCLUDES) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_30,code=\"sm_30,compute_30\" -o $@ -c $<
$(NVCC) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_30,code=\"sm_30,compute_30\" -o $@ -c $<
scrypt/nv_kernel.o: scrypt/nv_kernel.cu
$(NVCC) $(JANSSON_INCLUDES) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_30,code=\"sm_30,compute_30\" -o $@ -c $<
$(NVCC) -I. @CUDA_INCLUDES@ @CUDA_CFLAGS@ -gencode=arch=compute_30,code=\"sm_30,compute_30\" -o $@ -c $<
scrypt/titan_kernel.o: scrypt/titan_kernel.cu
$(NVCC) $(nvcc_FLAGS) -gencode=arch=compute_35,code=\"sm_35,compute_35\" -o $@ -c $<
+7 -33
View File
@@ -1,5 +1,5 @@
ccMiner release 1.6.5-C11 (Jul 2015) "Special for Chaincoin C11"
ccMiner release 1.6.4-tpruvot (May 2015) - "Multi-pools config"
---------------------------------------------------------------
***************************************************************
@@ -31,11 +31,9 @@ JackpotCoin
QuarkCoin family & AnimeCoin
TalkCoin
DarkCoin and other X11 coins
Chaincoin and Flaxscript (C11)
Saffroncoin blake (256 14-rounds)
BlakeCoin (256 8-rounds)
Qubit (Digibyte, ...)
Luffa (Joincoin)
Deep, Doom and Qubit
Keccak (Maxcoin)
Pentablake (Blake 512 x5)
1Coin Triple S
@@ -67,7 +65,6 @@ its command line interface and options.
anime use to mine Animecoin
blake use to mine Saffroncoin (Blake 256)
blakecoin use to mine Old Blake 256
c11/flax use to mine Chaincoin and Flax
deep use to mine Deepcoin
dmd-gr use to mine Diamond-Groestl
fresh use to mine Freshcoin
@@ -76,13 +73,14 @@ its command line interface and options.
heavy use to mine Heavycoin
jackpot use to mine Jackpotcoin
keccak use to mine Maxcoin
luffa use to mine Joincoin
luffa use to mine Doomcoin
lyra2 use to mine Vertcoin
mjollnir use to mine Mjollnircoin
myr-gr use to mine Myriad-Groest
neoscrypt use to mine FeatherCoin
nist5 use to mine TalkCoin
penta use to mine Joincoin / Pentablake
pluck use to mine Supcoin
quark use to mine Quarkcoin
qubit use to mine Qubit
scrypt use to mine Scrypt coins
@@ -95,7 +93,6 @@ its command line interface and options.
x14 use to mine X14Coin
x15 use to mine Halcyon
x17 use to mine X17
whirlpoolx use to mine Vanilla
zr5 use to mine ZiftrCoin
-d, --devices gives a comma separated list of CUDA device IDs
@@ -105,10 +102,8 @@ its command line interface and options.
-i, --intensity=N[,N] GPU threads per call 8-25 (2^N + F, default: 0=auto)
Decimals and multiple values are allowed for fine tuning
-f, --diff-factor Divide difficulty by this factor (default 1.0)
-m, --diff-multiplier Multiply difficulty by this value (default 1.0)
--vote=VOTE block reward vote (for HeavyCoin)
--trust-pool trust the max block reward vote (maxvote) sent by the pool
-f, --diff Divide difficulty by this factor (std is 1)
-v, --vote Heavycoin block vote (default: 512)
-o, --url=URL URL of mining server
-O, --userpass=U:P username:password pair for mining server
-u, --user=USERNAME username for mining server
@@ -137,9 +132,6 @@ its command line interface and options.
--max-temp=N Only mine if gpu temp is less than specified value
--max-rate=N[KMG] Only mine if net hashrate is less than specified value
--max-diff=N Only mine if net difficulty is less than specified value
--pstate=0 will force the Geforce 9xx to run in P0 P-State
--plimit=150W set the gpu power limit, allow multiple values for N cards
--keep-clocks prevent reset clocks and/or power limit on exit
-B, --background run the miner in the background
--benchmark run in offline benchmark mode
--cputest debug hashes from cpu algorithms
@@ -223,25 +215,7 @@ features.
>>> RELEASE HISTORY <<<
August 2015...
Restore WhirlpoolX algo
July 06th 2015 v1.6.5-C11
Nvml api power limits
Add chaincoin c11 algo (used by Flaxscript too)
Remove pluck algo
June 23th 2015 v1.6.5
Handle Ziftrcoin PoK solo mining
Basic compatibility with CUDA 7.0 (generally slower hashrate)
Show gpus vendor names on linux (windows test branch is pciutils)
Remove -v and -m short params specific to heavycoin
Add --diff-multiplier (-m) and rename --diff to --diff-factor (-f)
First steps to handle nvml application clocks and P0 on the GTX9xx
Various improvements on multipool and cmdline parameters
Optimize a bit qubit, deep, luffa, x11 and quark algos
May 26th 2015 v1.6.4
May 21th 2015 v1.6.4
Implement multi-pool support (failover and time rotate)
try "ccminer -c pools.conf" to test the sample config
Update the API to allow remote pool switching and pool stats
+15 -20
View File
@@ -8,7 +8,7 @@
* Software Foundation; either version 2 of the License, or (at your option)
* any later version. See COPYING for more details.
*/
#define APIVERSION "1.6"
#define APIVERSION "1.5"
#ifdef WIN32
# define _WINSOCK_DEPRECATED_NO_WARNINGS
@@ -124,7 +124,6 @@ static void gpustatus(int thr_id)
cgpu->gpu_temp = gpu_temp(cgpu);
cgpu->gpu_fan = (uint16_t) gpu_fanpercent(cgpu);
cgpu->gpu_fan_rpm = (uint16_t) gpu_fanrpm(cgpu);
cgpu->gpu_power = gpu_power(cgpu);
#endif
cuda_gpu_clocks(cgpu);
@@ -136,11 +135,10 @@ static void gpustatus(int thr_id)
card = device_name[gpuid];
snprintf(buf, sizeof(buf), "GPU=%d;BUS=%hd;CARD=%s;TEMP=%.1f;"
"POWER=%u;FAN=%hu;RPM=%hu;FREQ=%d;KHS=%.2f;HWF=%d;I=%.1f;THR=%u|",
gpuid, cgpu->gpu_bus, card, cgpu->gpu_temp,
cgpu->gpu_power, cgpu->gpu_fan, cgpu->gpu_fan_rpm,
cgpu->gpu_clock, cgpu->khashes,
snprintf(buf, sizeof(buf), "GPU=%d;BUS=%hd;CARD=%s;"
"TEMP=%.1f;FAN=%hu;RPM=%hu;FREQ=%d;KHS=%.2f;HWF=%d;I=%.1f;THR=%u|",
gpuid, cgpu->gpu_bus, card, cgpu->gpu_temp, cgpu->gpu_fan,
cgpu->gpu_fan_rpm, cgpu->gpu_clock, cgpu->khashes,
cgpu->hw_errors, cgpu->intensity, cgpu->throughput);
// append to buffer for multi gpus
@@ -171,9 +169,9 @@ static char *getsummary(char *params)
double accps, uptime = difftime(ts, startup);
uint32_t wait_time = 0, accepted_count = 0, rejected_count = 0;
for (int p = 0; p < num_pools; p++) {
wait_time += pools[p].wait_time;
accepted_count += pools[p].accepted_count;
rejected_count += pools[p].rejected_count;
wait_time += pools[cur_pooln].wait_time;
accepted_count += pools[cur_pooln].accepted_count;
rejected_count += pools[cur_pooln].rejected_count;
}
accps = (60.0 * accepted_count) / (uptime ? uptime : 1.0);
@@ -182,12 +180,12 @@ static char *getsummary(char *params)
*buffer = '\0';
sprintf(buffer, "NAME=%s;VER=%s;API=%s;"
"ALGO=%s;GPUS=%d;KHS=%.2f;ACC=%d;REJ=%d;"
"ACCMN=%.3f;DIFF=%.6f;NETKHS=%.0f;"
"ACCMN=%.3f;DIFF=%.6f;NETKHS=%.2f;"
"POOLS=%u;WAIT=%u;UPTIME=%.0f;TS=%u|",
PACKAGE_NAME, PACKAGE_VERSION, APIVERSION,
algo, active_gpus, (double)global_hashrate / 1000.,
accepted_count, rejected_count,
accps, net_diff > 1e-6 ? net_diff : stratum_diff, (double)net_hashrate / 1000.,
accps, net_diff > 0. ? net_diff : stratum_diff, (double)net_hashrate / 1000.,
num_pools, wait_time, uptime, (uint32_t) ts);
return buffer;
}
@@ -200,8 +198,7 @@ static char *getpoolnfo(char *params)
char *s = buffer;
char jobid[128] = { 0 };
char nonce[128] = { 0 };
int pooln = params ? atoi(params) % num_pools : cur_pooln;
struct pool_infos *p = &pools[pooln];
struct pool_infos *p = &pools[cur_pooln];
*s = '\0';
@@ -249,8 +246,7 @@ static void gpuhwinfos(int gpu_id)
cgpu->gpu_temp = gpu_temp(cgpu);
cgpu->gpu_fan = (uint16_t) gpu_fanpercent(cgpu);
cgpu->gpu_fan_rpm = (uint16_t) gpu_fanrpm(cgpu);
cgpu->gpu_pstate = (int16_t) gpu_pstate(cgpu);
cgpu->gpu_power = gpu_power(cgpu);
cgpu->gpu_pstate = gpu_pstate(cgpu);
gpu_info(cgpu);
#endif
@@ -258,18 +254,17 @@ static void gpuhwinfos(int gpu_id)
memset(pstate, 0, sizeof(pstate));
if (cgpu->gpu_pstate != -1)
snprintf(pstate, sizeof(pstate), "P%d", (int) cgpu->gpu_pstate);
snprintf(pstate, sizeof(pstate), "P%hu", cgpu->gpu_pstate);
card = device_name[gpu_id];
snprintf(buf, sizeof(buf), "GPU=%d;BUS=%hd;CARD=%s;SM=%u;MEM=%lu;"
"TEMP=%.1f;FAN=%hu;RPM=%hu;FREQ=%d;MEMFREQ=%d;PST=%s;POWER=%u;"
"TEMP=%.1f;FAN=%hu;RPM=%hu;FREQ=%d;MEMFREQ=%d;PST=%s;"
"VID=%hx;PID=%hx;NVML=%d;NVAPI=%d;SN=%s;BIOS=%s|",
gpu_id, cgpu->gpu_bus, card, cgpu->gpu_arch, cgpu->gpu_mem,
cgpu->gpu_temp, cgpu->gpu_fan, cgpu->gpu_fan_rpm,
cgpu->gpu_clock, cgpu->gpu_memclock,
pstate, cgpu->gpu_power,
cgpu->gpu_vid, cgpu->gpu_pid, cgpu->nvml_id, cgpu->nvapi_id,
pstate, cgpu->gpu_vid, cgpu->gpu_pid, cgpu->nvml_id, cgpu->nvapi_id,
cgpu->gpu_sn, cgpu->gpu_desc);
strcat(buffer, buf);
-4
View File
@@ -62,7 +62,6 @@ function translateField($key)
$intl['TEMP'] = 'T°c';
$intl['FAN'] = 'Fan %';
$intl['FREQ'] = 'Freq.';
$intl['POWER'] = 'Power';
$intl['PST'] = 'P-State';
// pool infos
@@ -97,9 +96,6 @@ function translateValue($key,$val,$data=array())
case 'FREQ':
$val = sprintf("%d MHz", round(floatval($val)/1000.0));
break;
case 'POWER':
$val = sprintf("%d W", round(floatval($val)/1000.0));
break;
case 'TS':
$val = strftime("%H:%M:%S", (int) $val);
break;
-26
View File
@@ -1,26 +0,0 @@
/**
* Wrapper to OpenSSL BIGNUM used by net diff (nBits)
*/
#include <stdio.h>
#include "uint256.h"
#include "bignum.hpp"
#include "miner.h" // hex2bin
extern "C" double bn_convert_nbits(const uint32_t nBits)
{
uint256 bn = CBigNum().SetCompact(nBits).getuint256();
return bn.getdouble();
}
// copy the big number to 32-bytes uchar
extern "C" void bn_nbits_to_uchar(const uint32_t nBits, unsigned char *target)
{
char buff[65];
uint256 bn = CBigNum().SetCompact(nBits).getuint256();
snprintf(buff, 65, "%s\n", bn.ToString().c_str()); buff[64] = '\0';
hex2bin(target, buff, 32);
}
-602
View File
@@ -1,602 +0,0 @@
// Copyright (c) 2009-2010 Satoshi Nakamoto
// Copyright (c) 2009-2012 The Bitcoin developers
// Distributed under the MIT/X11 software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#ifndef BITCOIN_BIGNUM_H
#define BITCOIN_BIGNUM_H
#ifndef PROTOCOL_VERSION
#define PROTOCOL_VERSION 70001
#endif
#include <map>
#include <limits>
#include <algorithm>
#include <stdexcept>
#include <vector>
#include "uint256.h"
#include <openssl/bn.h>
#include "serialize.hpp"
using namespace std;
/** Errors thrown by the bignum class */
class bignum_error : public std::runtime_error
{
public:
explicit bignum_error(const std::string& str) : std::runtime_error(str) {}
};
/** RAII encapsulated BN_CTX (OpenSSL bignum context) */
class CAutoBN_CTX
{
protected:
BN_CTX* pctx;
BN_CTX* operator=(BN_CTX* pnew) { return pctx = pnew; }
public:
CAutoBN_CTX()
{
pctx = BN_CTX_new();
if (pctx == NULL)
throw bignum_error("CAutoBN_CTX : BN_CTX_new() returned NULL");
}
~CAutoBN_CTX()
{
if (pctx != NULL)
BN_CTX_free(pctx);
}
operator BN_CTX*() { return pctx; }
BN_CTX& operator*() { return *pctx; }
BN_CTX** operator&() { return &pctx; }
bool operator!() { return (pctx == NULL); }
};
/** C++ wrapper for BIGNUM (OpenSSL bignum) */
class CBigNum : public BIGNUM
{
public:
CBigNum()
{
BN_init(this);
}
CBigNum(const CBigNum& b)
{
BN_init(this);
if (!BN_copy(this, &b))
{
BN_clear_free(this);
throw bignum_error("CBigNum::CBigNum(const CBigNum&) : BN_copy failed");
}
}
CBigNum& operator=(const CBigNum& b)
{
if (!BN_copy(this, &b))
throw bignum_error("CBigNum::operator= : BN_copy failed");
return (*this);
}
~CBigNum()
{
BN_clear_free(this);
}
//CBigNum(char n) is not portable. Use 'signed char' or 'unsigned char'.
CBigNum(signed char n) { BN_init(this); if (n >= 0) setulong(n); else setint64(n); }
CBigNum(short n) { BN_init(this); if (n >= 0) setulong(n); else setint64(n); }
CBigNum(int n) { BN_init(this); if (n >= 0) setulong(n); else setint64(n); }
CBigNum(long n) { BN_init(this); if (n >= 0) setulong(n); else setint64(n); }
CBigNum(int64 n) { BN_init(this); setint64(n); }
CBigNum(unsigned char n) { BN_init(this); setulong(n); }
CBigNum(unsigned short n) { BN_init(this); setulong(n); }
CBigNum(unsigned int n) { BN_init(this); setulong(n); }
CBigNum(unsigned long n) { BN_init(this); setulong(n); }
CBigNum(uint64 n) { BN_init(this); setuint64(n); }
explicit CBigNum(uint256 n) { BN_init(this); setuint256(n); }
explicit CBigNum(const std::vector<unsigned char>& vch)
{
BN_init(this);
setvch(vch);
}
void setulong(unsigned long n)
{
if (!BN_set_word(this, n))
throw bignum_error("CBigNum conversion from unsigned long : BN_set_word failed");
}
unsigned long getulong() const
{
return BN_get_word(this);
}
unsigned int getuint() const
{
return BN_get_word(this);
}
int getint() const
{
unsigned long n = BN_get_word(this);
if (!BN_is_negative(this))
return (n > (unsigned long)std::numeric_limits<int>::max() ? std::numeric_limits<int>::max() : n);
else
return (n > (unsigned long)std::numeric_limits<int>::max() ? std::numeric_limits<int>::min() : -(int)n);
}
void setint64(int64 sn)
{
unsigned char pch[sizeof(sn) + 6];
unsigned char* p = pch + 4;
bool fNegative;
uint64 n;
if (sn < (int64)0)
{
// Since the minimum signed integer cannot be represented as positive so long as its type is signed,
// and it's not well-defined what happens if you make it unsigned before negating it,
// we instead increment the negative integer by 1, convert it, then increment the (now positive) unsigned integer by 1 to compensate
n = -(sn + 1);
++n;
fNegative = true;
} else {
n = sn;
fNegative = false;
}
bool fLeadingZeroes = true;
for (int i = 0; i < 8; i++)
{
unsigned char c = (n >> 56) & 0xff;
n <<= 8;
if (fLeadingZeroes)
{
if (c == 0)
continue;
if (c & 0x80)
*p++ = (fNegative ? 0x80 : 0);
else if (fNegative)
c |= 0x80;
fLeadingZeroes = false;
}
*p++ = c;
}
unsigned int nSize = p - (pch + 4);
pch[0] = (nSize >> 24) & 0xff;
pch[1] = (nSize >> 16) & 0xff;
pch[2] = (nSize >> 8) & 0xff;
pch[3] = (nSize) & 0xff;
BN_mpi2bn(pch, p - pch, this);
}
void setuint64(uint64 n)
{
unsigned char pch[sizeof(n) + 6];
unsigned char* p = pch + 4;
bool fLeadingZeroes = true;
for (int i = 0; i < 8; i++)
{
unsigned char c = (n >> 56) & 0xff;
n <<= 8;
if (fLeadingZeroes)
{
if (c == 0)
continue;
if (c & 0x80)
*p++ = 0;
fLeadingZeroes = false;
}
*p++ = c;
}
unsigned int nSize = p - (pch + 4);
pch[0] = (nSize >> 24) & 0xff;
pch[1] = (nSize >> 16) & 0xff;
pch[2] = (nSize >> 8) & 0xff;
pch[3] = (nSize) & 0xff;
BN_mpi2bn(pch, p - pch, this);
}
void setuint256(uint256 n)
{
unsigned char pch[sizeof(n) + 6];
unsigned char* p = pch + 4;
bool fLeadingZeroes = true;
unsigned char* pbegin = (unsigned char*)&n;
unsigned char* psrc = pbegin + sizeof(n);
while (psrc != pbegin)
{
unsigned char c = *(--psrc);
if (fLeadingZeroes)
{
if (c == 0)
continue;
if (c & 0x80)
*p++ = 0;
fLeadingZeroes = false;
}
*p++ = c;
}
unsigned int nSize = p - (pch + 4);
pch[0] = (nSize >> 24) & 0xff;
pch[1] = (nSize >> 16) & 0xff;
pch[2] = (nSize >> 8) & 0xff;
pch[3] = (nSize >> 0) & 0xff;
BN_mpi2bn(pch, p - pch, this);
}
uint256 getuint256() const
{
unsigned int nSize = BN_bn2mpi(this, NULL);
if (nSize < 4)
return 0;
std::vector<unsigned char> vch(nSize);
BN_bn2mpi(this, &vch[0]);
if (vch.size() > 4)
vch[4] &= 0x7f;
uint256 n = 0;
for (unsigned int i = 0, j = vch.size()-1; i < sizeof(n) && j >= 4; i++, j--)
((unsigned char*)&n)[i] = vch[j];
return n;
}
void setvch(const std::vector<unsigned char>& vch)
{
std::vector<unsigned char> vch2(vch.size() + 4);
unsigned int nSize = vch.size();
// BIGNUM's byte stream format expects 4 bytes of
// big endian size data info at the front
vch2[0] = (nSize >> 24) & 0xff;
vch2[1] = (nSize >> 16) & 0xff;
vch2[2] = (nSize >> 8) & 0xff;
vch2[3] = (nSize >> 0) & 0xff;
// swap data to big endian
reverse_copy(vch.begin(), vch.end(), vch2.begin() + 4);
BN_mpi2bn(&vch2[0], vch2.size(), this);
}
std::vector<unsigned char> getvch() const
{
unsigned int nSize = BN_bn2mpi(this, NULL);
if (nSize <= 4)
return std::vector<unsigned char>();
std::vector<unsigned char> vch(nSize);
BN_bn2mpi(this, &vch[0]);
vch.erase(vch.begin(), vch.begin() + 4);
reverse(vch.begin(), vch.end());
return vch;
}
// The "compact" format is a representation of a whole
// number N using an unsigned 32bit number similar to a
// floating point format.
// The most significant 8 bits are the unsigned exponent of base 256.
// This exponent can be thought of as "number of bytes of N".
// The lower 23 bits are the mantissa.
// Bit number 24 (0x800000) represents the sign of N.
// N = (-1^sign) * mantissa * 256^(exponent-3)
//
// Satoshi's original implementation used BN_bn2mpi() and BN_mpi2bn().
// MPI uses the most significant bit of the first byte as sign.
// Thus 0x1234560000 is compact (0x05123456)
// and 0xc0de000000 is compact (0x0600c0de)
// (0x05c0de00) would be -0x40de000000
//
// Bitcoin only uses this "compact" format for encoding difficulty
// targets, which are unsigned 256bit quantities. Thus, all the
// complexities of the sign bit and using base 256 are probably an
// implementation accident.
//
// This implementation directly uses shifts instead of going
// through an intermediate MPI representation.
CBigNum& SetCompact(unsigned int nCompact)
{
unsigned int nSize = nCompact >> 24;
bool fNegative =(nCompact & 0x00800000) != 0;
unsigned int nWord = nCompact & 0x007fffff;
if (nSize <= 3)
{
nWord >>= 8*(3-nSize);
BN_set_word(this, nWord);
}
else
{
BN_set_word(this, nWord);
BN_lshift(this, this, 8*(nSize-3));
}
BN_set_negative(this, fNegative);
return *this;
}
unsigned int GetCompact() const
{
unsigned int nSize = BN_num_bytes(this);
unsigned int nCompact = 0;
if (nSize <= 3)
nCompact = BN_get_word(this) << 8*(3-nSize);
else
{
CBigNum bn;
BN_rshift(&bn, this, 8*(nSize-3));
nCompact = BN_get_word(&bn);
}
// The 0x00800000 bit denotes the sign.
// Thus, if it is already set, divide the mantissa by 256 and increase the exponent.
if (nCompact & 0x00800000)
{
nCompact >>= 8;
nSize++;
}
nCompact |= nSize << 24;
nCompact |= (BN_is_negative(this) ? 0x00800000 : 0);
return nCompact;
}
void SetHex(const std::string& str)
{
// skip 0x
const char* psz = str.c_str();
while (isspace(*psz))
psz++;
bool fNegative = false;
if (*psz == '-')
{
fNegative = true;
psz++;
}
if (psz[0] == '0' && tolower(psz[1]) == 'x')
psz += 2;
while (isspace(*psz))
psz++;
// hex string to bignum
static const signed char phexdigit[256] = { 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,1,2,3,4,5,6,7,8,9,0,0,0,0,0,0, 0,0xa,0xb,0xc,0xd,0xe,0xf,0,0,0,0,0,0,0,0,0, 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0, 0,0xa,0xb,0xc,0xd,0xe,0xf,0,0,0,0,0,0,0,0,0 };
*this = 0;
while (isxdigit(*psz))
{
*this <<= 4;
int n = phexdigit[(unsigned char)*psz++];
*this += n;
}
if (fNegative)
*this = 0 - *this;
}
std::string ToString(int nBase=10) const
{
CAutoBN_CTX pctx;
CBigNum bnBase = nBase;
CBigNum bn0 = 0;
std::string str;
CBigNum bn = *this;
BN_set_negative(&bn, false);
CBigNum dv;
CBigNum rem;
if (BN_cmp(&bn, &bn0) == 0)
return "0";
while (BN_cmp(&bn, &bn0) > 0)
{
if (!BN_div(&dv, &rem, &bn, &bnBase, pctx))
throw bignum_error("CBigNum::ToString() : BN_div failed");
bn = dv;
unsigned int c = rem.getulong();
str += "0123456789abcdef"[c];
}
if (BN_is_negative(this))
str += "-";
reverse(str.begin(), str.end());
return str;
}
std::string GetHex() const
{
return ToString(16);
}
unsigned int GetSerializeSize(int nType=0, int nVersion=PROTOCOL_VERSION) const
{
return ::GetSerializeSize(getvch(), nType, nVersion);
}
template<typename Stream>
void Serialize(Stream& s, int nType=0, int nVersion=PROTOCOL_VERSION) const
{
::Serialize(s, getvch(), nType, nVersion);
}
template<typename Stream>
void Unserialize(Stream& s, int nType=0, int nVersion=PROTOCOL_VERSION)
{
std::vector<unsigned char> vch;
::Unserialize(s, vch, nType, nVersion);
setvch(vch);
}
bool operator!() const
{
return BN_is_zero(this);
}
CBigNum& operator+=(const CBigNum& b)
{
if (!BN_add(this, this, &b))
throw bignum_error("CBigNum::operator+= : BN_add failed");
return *this;
}
CBigNum& operator-=(const CBigNum& b)
{
*this = *this - b;
return *this;
}
CBigNum& operator*=(const CBigNum& b)
{
CAutoBN_CTX pctx;
if (!BN_mul(this, this, &b, pctx))
throw bignum_error("CBigNum::operator*= : BN_mul failed");
return *this;
}
CBigNum& operator/=(const CBigNum& b)
{
*this = *this / b;
return *this;
}
CBigNum& operator%=(const CBigNum& b)
{
*this = *this % b;
return *this;
}
CBigNum& operator<<=(unsigned int shift)
{
if (!BN_lshift(this, this, shift))
throw bignum_error("CBigNum:operator<<= : BN_lshift failed");
return *this;
}
CBigNum& operator>>=(unsigned int shift)
{
// Note: BN_rshift segfaults on 64-bit if 2^shift is greater than the number
// if built on ubuntu 9.04 or 9.10, probably depends on version of OpenSSL
CBigNum a = 1;
a <<= shift;
if (BN_cmp(&a, this) > 0)
{
*this = 0;
return *this;
}
if (!BN_rshift(this, this, shift))
throw bignum_error("CBigNum:operator>>= : BN_rshift failed");
return *this;
}
CBigNum& operator++()
{
// prefix operator
if (!BN_add(this, this, BN_value_one()))
throw bignum_error("CBigNum::operator++ : BN_add failed");
return *this;
}
const CBigNum operator++(int)
{
// postfix operator
const CBigNum ret = *this;
++(*this);
return ret;
}
CBigNum& operator--()
{
// prefix operator
CBigNum r;
if (!BN_sub(&r, this, BN_value_one()))
throw bignum_error("CBigNum::operator-- : BN_sub failed");
*this = r;
return *this;
}
const CBigNum operator--(int)
{
// postfix operator
const CBigNum ret = *this;
--(*this);
return ret;
}
friend inline const CBigNum operator-(const CBigNum& a, const CBigNum& b);
friend inline const CBigNum operator/(const CBigNum& a, const CBigNum& b);
friend inline const CBigNum operator%(const CBigNum& a, const CBigNum& b);
};
inline const CBigNum operator+(const CBigNum& a, const CBigNum& b)
{
CBigNum r;
if (!BN_add(&r, &a, &b))
throw bignum_error("CBigNum::operator+ : BN_add failed");
return r;
}
inline const CBigNum operator-(const CBigNum& a, const CBigNum& b)
{
CBigNum r;
if (!BN_sub(&r, &a, &b))
throw bignum_error("CBigNum::operator- : BN_sub failed");
return r;
}
inline const CBigNum operator-(const CBigNum& a)
{
CBigNum r(a);
BN_set_negative(&r, !BN_is_negative(&r));
return r;
}
inline const CBigNum operator*(const CBigNum& a, const CBigNum& b)
{
CAutoBN_CTX pctx;
CBigNum r;
if (!BN_mul(&r, &a, &b, pctx))
throw bignum_error("CBigNum::operator* : BN_mul failed");
return r;
}
inline const CBigNum operator/(const CBigNum& a, const CBigNum& b)
{
CAutoBN_CTX pctx;
CBigNum r;
if (!BN_div(&r, NULL, &a, &b, pctx))
throw bignum_error("CBigNum::operator/ : BN_div failed");
return r;
}
inline const CBigNum operator%(const CBigNum& a, const CBigNum& b)
{
CAutoBN_CTX pctx;
CBigNum r;
if (!BN_mod(&r, &a, &b, pctx))
throw bignum_error("CBigNum::operator% : BN_div failed");
return r;
}
inline const CBigNum operator<<(const CBigNum& a, unsigned int shift)
{
CBigNum r;
if (!BN_lshift(&r, &a, shift))
throw bignum_error("CBigNum:operator<< : BN_lshift failed");
return r;
}
inline const CBigNum operator>>(const CBigNum& a, unsigned int shift)
{
CBigNum r = a;
r >>= shift;
return r;
}
inline bool operator==(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) == 0); }
inline bool operator!=(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) != 0); }
inline bool operator<=(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) <= 0); }
inline bool operator>=(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) >= 0); }
inline bool operator<(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) < 0); }
inline bool operator>(const CBigNum& a, const CBigNum& b) { return (BN_cmp(&a, &b) > 0); }
#endif
+417 -269
View File
File diff suppressed because it is too large. Load diff
+50 -34
View File
@@ -23,22 +23,34 @@
<RootNamespace>ccminer</RootNamespace>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.Default.props" />
<PropertyGroup Condition="'$(Configuration)'=='Release'" Label="Configuration">
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v120</PlatformToolset>
<CLRSupport>false</CLRSupport>
<LinkIncremental>false</LinkIncremental>
<WholeProgramOptimization>true</WholeProgramOptimization>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)'=='Debug'" Label="Configuration">
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v120</PlatformToolset>
<CLRSupport>false</CLRSupport>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>false</UseDebugLibraries>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v120</PlatformToolset>
<CLRSupport>false</CLRSupport>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'" Label="Configuration">
<ConfigurationType>Application</ConfigurationType>
<UseDebugLibraries>true</UseDebugLibraries>
<CharacterSet>MultiByte</CharacterSet>
<PlatformToolset>v120</PlatformToolset>
<CLRSupport>false</CLRSupport>
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<Import Project="$(VCTargetsPath)\Microsoft.Cpp.props" />
<ImportGroup Label="ExtensionSettings">
@@ -56,6 +68,19 @@
<ImportGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'" Label="PropertySheets">
<Import Project="$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props" Condition="exists('$(UserRootDir)\Microsoft.Cpp.$(Platform).user.props')" Label="LocalAppDataPlatform" />
</ImportGroup>
<PropertyGroup Label="UserMacros" />
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<LinkIncremental>true</LinkIncremental>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|Win32'">
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<PropertyGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
<LinkIncremental>false</LinkIncremental>
</PropertyGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">
<ClCompile>
<WarningLevel>Level3</WarningLevel>
@@ -66,7 +91,6 @@
<AdditionalIncludeDirectories>.;compat;compat\curl-for-windows\curl\include;compat\jansson;compat\getopt;compat\pthreads;compat\curl-for-windows\openssl\openssl\include;compat\curl-for-windows\zlib;%(AdditionalIncludeDirectories);$(CudaToolkitIncludeDir)</AdditionalIncludeDirectories>
<EnableParallelCodeGeneration>true</EnableParallelCodeGeneration>
<MultiProcessorCompilation>true</MultiProcessorCompilation>
<OpenMPSupport>true</OpenMPSupport>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
@@ -84,7 +108,6 @@
<PtxAsOptionV>true</PtxAsOptionV>
<Keep>true</Keep>
<CodeGeneration>compute_50,sm_50</CodeGeneration>
<Include>$(NVTOOLSEXT_PATH)\include;..\..\..\Common\C99</Include>
</CudaCompile>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Debug|x64'">
@@ -98,7 +121,6 @@
<StructMemberAlignment>8Bytes</StructMemberAlignment>
<EnableParallelCodeGeneration>true</EnableParallelCodeGeneration>
<MultiProcessorCompilation>true</MultiProcessorCompilation>
<OpenMPSupport>true</OpenMPSupport>
</ClCompile>
<Link>
<GenerateDebugInformation>true</GenerateDebugInformation>
@@ -137,7 +159,7 @@
<AdditionalIncludeDirectories>.;compat;compat\curl-for-windows\curl\include;compat\jansson;compat\getopt;compat\pthreads;compat\curl-for-windows\openssl\openssl\include;compat\curl-for-windows\zlib;%(AdditionalIncludeDirectories);$(CudaToolkitIncludeDir)</AdditionalIncludeDirectories>
<EnableParallelCodeGeneration>true</EnableParallelCodeGeneration>
<MultiProcessorCompilation>true</MultiProcessorCompilation>
<OpenMPSupport>true</OpenMPSupport>
<ExceptionHandling>SyncCThrow</ExceptionHandling>
</ClCompile>
<Link>
<GenerateDebugInformation>false</GenerateDebugInformation>
@@ -157,13 +179,17 @@
<MaxRegCount>80</MaxRegCount>
<PtxAsOptionV>true</PtxAsOptionV>
<Keep>true</Keep>
<CodeGeneration>compute_50,sm_50;compute_52,sm_52;compute_30,sm_30;compute_35,sm_35</CodeGeneration>
<CodeGeneration>compute_35,sm_35;compute_50,sm_50;compute_52,sm_52</CodeGeneration>
<AdditionalOptions>--ptxas-options="-O2" %(AdditionalOptions)</AdditionalOptions>
<Optimization>O2</Optimization>
<Defines>
</Defines>
<Optimization>O3</Optimization>
</CudaCompile>
<CudaLink>
<GPUDebugInfo>false</GPUDebugInfo>
<Optimization>O3</Optimization>
</CudaLink>
<CudaLink>
<Optimization>O2</Optimization>
</CudaLink>
</ItemDefinitionGroup>
<ItemDefinitionGroup Condition="'$(Configuration)|$(Platform)'=='Release|x64'">
@@ -172,9 +198,7 @@
<Optimization>MaxSpeed</Optimization>
<RuntimeLibrary>MultiThreaded</RuntimeLibrary>
<FavorSizeOrSpeed>Speed</FavorSizeOrSpeed>
<EnableEnhancedInstructionSet>
</EnableEnhancedInstructionSet>
<OpenMPSupport>true</OpenMPSupport>
<EnableEnhancedInstructionSet>AdvancedVectorExtensions</EnableEnhancedInstructionSet>
<CompileAsManaged>false</CompileAsManaged>
<FunctionLevelLinking>true</FunctionLevelLinking>
<IntrinsicFunctions>true</IntrinsicFunctions>
@@ -198,14 +222,13 @@
<CInterleavedPTX>false</CInterleavedPTX>
<MaxRegCount>80</MaxRegCount>
<PtxAsOptionV>true</PtxAsOptionV>
<Keep>true</Keep>
<Keep>false</Keep>
<CodeGeneration>compute_35,sm_35;compute_50,sm_50;compute_52,sm_52</CodeGeneration>
<Include>$(NVTOOLSEXT_PATH)\include;..\..\..\Common\C99</Include>
<Optimization>O3</Optimization>
<Include>
</Include>
<TargetMachinePlatform>64</TargetMachinePlatform>
</CudaCompile>
<CudaLink>
<Optimization>O3</Optimization>
<PerformDeviceLink>false</PerformDeviceLink>
</CudaLink>
</ItemDefinitionGroup>
@@ -227,10 +250,9 @@
<TreatWChar_tAsBuiltInType>false</TreatWChar_tAsBuiltInType>
<Optimization Condition="'$(Configuration)'=='Release'">Full</Optimization>
</ClCompile>
<ClCompile Include="pools.cpp" />
<ClCompile Include="scrypt-jane.cpp" />
<ClCompile Include="scrypt.cpp" />
<ClCompile Include="util.cpp" />
<ClCompile Include="bignum.cpp" />
<ClInclude Include="bignum.hpp" />
<ClCompile Include="fuguecoin.cpp" />
<ClCompile Include="groestlcoin.cpp" />
<ClCompile Include="hashlog.cpp" />
@@ -248,8 +270,6 @@
<ClCompile Include="neoscrypt\neoscrypt-cpu.c" />
<ClInclude Include="neoscrypt\cuda_vectors.h" />
<CudaCompile Include="neoscrypt\cuda_neoscrypt.cu" />
<ClCompile Include="scrypt-jane.cpp" />
<ClCompile Include="scrypt.cpp" />
<ClCompile Include="skein2.cpp" />
<ClCompile Include="sph\aes_helper.c" />
<ClCompile Include="sph\blake.c" />
@@ -421,6 +441,8 @@
<AdditionalOptions Condition="'$(Configuration)'=='Release'">-Xptxas "-abi=yes" %(AdditionalOptions)</AdditionalOptions>
<AdditionalOptions Condition="'$(Configuration)'=='Debug'">-Xptxas "-abi=yes" %(AdditionalOptions)</AdditionalOptions>
</CudaCompile>
<CudaCompile Include="pluck\pluck.cu" />
<CudaCompile Include="pluck\cuda_pluck.cu" />
<CudaCompile Include="quark\cuda_quark_groestl512.cu">
</CudaCompile>
<CudaCompile Include="quark\cuda_quark_keccak512.cu">
@@ -432,7 +454,7 @@
</CudaCompile>
<CudaCompile Include="qubit\deep.cu">
</CudaCompile>
<CudaCompile Include="qubit\luffa.cu">
<CudaCompile Include="qubit\doom.cu">
</CudaCompile>
<CudaCompile Include="qubit\qubit.cu">
</CudaCompile>
@@ -454,7 +476,7 @@
</CudaCompile>
<CudaCompile Include="x11\cuda_x11_luffa512.cu" />
<CudaCompile Include="x11\cuda_x11_luffa512_Cubehash.cu">
<MaxRegCount>76</MaxRegCount>
<MaxRegCount>80</MaxRegCount>
</CudaCompile>
<CudaCompile Include="x11\cuda_x11_shavite512.cu">
<MaxRegCount>128</MaxRegCount>
@@ -462,8 +484,8 @@
<CudaCompile Include="x11\cuda_x11_simd512.cu">
<MaxRegCount>64</MaxRegCount>
</CudaCompile>
<CudaCompile Include="x11\c11.cu" />
<CudaCompile Include="x11\fresh.cu" />
<CudaCompile Include="x11\fresh.cu">
</CudaCompile>
<CudaCompile Include="x11\s3.cu" />
<CudaCompile Include="x11\simd_functions.cu">
<ExcludedFromBuild>true</ExcludedFromBuild>
@@ -477,15 +499,9 @@
</CudaCompile>
<CudaCompile Include="x13\x13.cu">
</CudaCompile>
<CudaCompile Include="x15\x14.cu">
</CudaCompile>
<CudaCompile Include="x15\cuda_x14_shabal512.cu">
</CudaCompile>
<CudaCompile Include="x15\x15.cu">
</CudaCompile>
<CudaCompile Include="x15\cuda_whirlpoolx.cu" />
<CudaCompile Include="x15\whirlpoolx.cu" />
<CudaCompile Include="x17\x17.cu">
<CudaCompile Include="x15\cuda_x15_whirlpool.cu">
</CudaCompile>
<CudaCompile Include="x17\cuda_x17_haval512.cu">
</CudaCompile>
+7 -19
View File
@@ -111,9 +111,6 @@
<ClCompile Include="ccminer.cpp">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="pools.cpp">
<Filter>Source Files</Filter>
</ClCompile>
<ClCompile Include="crc32.c">
<Filter>Source Files</Filter>
</ClCompile>
@@ -246,9 +243,6 @@
<ClCompile Include="hefty1.c">
<Filter>Source Files\CUDA\heavy</Filter>
</ClCompile>
<ClCompile Include="bignum.cpp">
<Filter>Source Files</Filter>
</ClCompile>
</ItemGroup>
<ItemGroup>
<ClInclude Include="compat.h">
@@ -410,9 +404,6 @@
<ClInclude Include="hefty1.h">
<Filter>Source Files\CUDA\heavy</Filter>
</ClInclude>
<ClInclude Include="bignum.hpp">
<Filter>Header Files</Filter>
</ClInclude>
</ItemGroup>
<ItemGroup>
<CudaCompile Include="cuda.cpp">
@@ -433,6 +424,12 @@
<CudaCompile Include="JHA\cuda_jha_compactionTest.cu">
<Filter>Source Files\CUDA\JHA</Filter>
</CudaCompile>
<CudaCompile Include="pluck\pluck.cu">
<Filter>Source Files\CUDA</Filter>
</CudaCompile>
<CudaCompile Include="pluck\cuda_pluck.cu">
<Filter>Source Files\CUDA</Filter>
</CudaCompile>
<CudaCompile Include="quark\cuda_jh512.cu">
<Filter>Source Files\CUDA\quark</Filter>
</CudaCompile>
@@ -484,7 +481,7 @@
<CudaCompile Include="qubit\deep.cu">
<Filter>Source Files\CUDA\qubit</Filter>
</CudaCompile>
<CudaCompile Include="qubit\luffa.cu">
<CudaCompile Include="qubit\doom.cu">
<Filter>Source Files\CUDA\qubit</Filter>
</CudaCompile>
<CudaCompile Include="qubit\qubit.cu">
@@ -517,9 +514,6 @@
<CudaCompile Include="x11\cuda_x11_simd512.cu">
<Filter>Source Files\CUDA\x11</Filter>
</CudaCompile>
<CudaCompile Include="x11\c11.cu">
<Filter>Source Files\CUDA\x11</Filter>
</CudaCompile>
<CudaCompile Include="x11\fresh.cu">
<Filter>Source Files\CUDA\x11</Filter>
</CudaCompile>
@@ -553,12 +547,6 @@
<CudaCompile Include="x15\cuda_x15_whirlpool.cu">
<Filter>Source Files\CUDA\x15</Filter>
</CudaCompile>
<CudaCompile Include="x15\cuda_whirlpoolx.cu">
<Filter>Source Files\CUDA\x15</Filter>
</CudaCompile>
<CudaCompile Include="x15\whirlpoolx.cu">
<Filter>Source Files\CUDA\x15</Filter>
</CudaCompile>
<CudaCompile Include="x17\cuda_x17_haval512.cu">
<Filter>Source Files\CUDA\x17</Filter>
</CudaCompile>
@@ -1,902 +0,0 @@
/* crypto/bn/bn.h */
/* Copyright (C) 1995-1997 Eric Young (eay@cryptsoft.com)
* All rights reserved.
*
* This package is an SSL implementation written
* by Eric Young (eay@cryptsoft.com).
* The implementation was written so as to conform with Netscapes SSL.
*
* This library is free for commercial and non-commercial use as long as
* the following conditions are aheared to. The following conditions
* apply to all code found in this distribution, be it the RC4, RSA,
* lhash, DES, etc., code; not just the SSL code. The SSL documentation
* included with this distribution is covered by the same copyright terms
* except that the holder is Tim Hudson (tjh@cryptsoft.com).
*
* Copyright remains Eric Young's, and as such any Copyright notices in
* the code are not to be removed.
* If this package is used in a product, Eric Young should be given attribution
* as the author of the parts of the library used.
* This can be in the form of a textual message at program startup or
* in documentation (online or textual) provided with the package.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* "This product includes cryptographic software written by
* Eric Young (eay@cryptsoft.com)"
* The word 'cryptographic' can be left out if the rouines from the library
* being used are not cryptographic related :-).
* 4. If you include any Windows specific code (or a derivative thereof) from
* the apps directory (application code) you must include an acknowledgement:
* "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
*
* THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* The licence and distribution terms for any publically available version or
* derivative of this code cannot be changed. i.e. this code cannot simply be
* copied and put under another distribution licence
* [including the GNU Public Licence.]
*/
/* ====================================================================
* Copyright (c) 1998-2006 The OpenSSL Project. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* 3. All advertising materials mentioning features or use of this
* software must display the following acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
*
* 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
* endorse or promote products derived from this software without
* prior written permission. For written permission, please contact
* openssl-core@openssl.org.
*
* 5. Products derived from this software may not be called "OpenSSL"
* nor may "OpenSSL" appear in their names without prior written
* permission of the OpenSSL Project.
*
* 6. Redistributions of any form whatsoever must retain the following
* acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit (http://www.openssl.org/)"
*
* THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
* EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
* ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
* OF THE POSSIBILITY OF SUCH DAMAGE.
* ====================================================================
*
* This product includes cryptographic software written by Eric Young
* (eay@cryptsoft.com). This product includes software written by Tim
* Hudson (tjh@cryptsoft.com).
*
*/
/* ====================================================================
* Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
*
* Portions of the attached software ("Contribution") are developed by
* SUN MICROSYSTEMS, INC., and are contributed to the OpenSSL project.
*
* The Contribution is licensed pursuant to the Eric Young open source
* license provided above.
*
* The binary polynomial arithmetic software is originally written by
* Sheueling Chang Shantz and Douglas Stebila of Sun Microsystems Laboratories.
*
*/
#ifndef HEADER_BN_H
#define HEADER_BN_H
#include <openssl/e_os2.h>
#ifndef OPENSSL_NO_FP_API
#include <stdio.h> /* FILE */
#endif
#include <openssl/ossl_typ.h>
#include <openssl/crypto.h>
#ifdef __cplusplus
extern "C" {
#endif
/* These preprocessor symbols control various aspects of the bignum headers and
* library code. They're not defined by any "normal" configuration, as they are
* intended for development and testing purposes. NB: defining all three can be
* useful for debugging application code as well as openssl itself.
*
* BN_DEBUG - turn on various debugging alterations to the bignum code
* BN_DEBUG_RAND - uses random poisoning of unused words to trip up
* mismanagement of bignum internals. You must also define BN_DEBUG.
*/
/* #define BN_DEBUG */
/* #define BN_DEBUG_RAND */
#ifndef OPENSSL_SMALL_FOOTPRINT
#define BN_MUL_COMBA
#define BN_SQR_COMBA
#define BN_RECURSION
#endif
/* This next option uses the C libraries (2 word)/(1 word) function.
* If it is not defined, I use my C version (which is slower).
* The reason for this flag is that when the particular C compiler
* library routine is used, and the library is linked with a different
* compiler, the library is missing. This mostly happens when the
* library is built with gcc and then linked using normal cc. This would
* be a common occurrence because gcc normally produces code that is
* 2 times faster than system compilers for the big number stuff.
* For machines with only one compiler (or shared libraries), this should
* be on. Again this in only really a problem on machines
* using "long long's", are 32bit, and are not using my assembler code. */
#if defined(OPENSSL_SYS_MSDOS) || defined(OPENSSL_SYS_WINDOWS) || \
defined(OPENSSL_SYS_WIN32) || defined(linux)
# ifndef BN_DIV2W
# define BN_DIV2W
# endif
#endif
/* assuming long is 64bit - this is the DEC Alpha
* unsigned long long is only 64 bits :-(, don't define
* BN_LLONG for the DEC Alpha */
#ifdef SIXTY_FOUR_BIT_LONG
#define BN_ULLONG unsigned long long
#define BN_ULONG unsigned long
#define BN_LONG long
#define BN_BITS 128
#define BN_BYTES 8
#define BN_BITS2 64
#define BN_BITS4 32
#define BN_MASK (0xffffffffffffffffffffffffffffffffLL)
#define BN_MASK2 (0xffffffffffffffffL)
#define BN_MASK2l (0xffffffffL)
#define BN_MASK2h (0xffffffff00000000L)
#define BN_MASK2h1 (0xffffffff80000000L)
#define BN_TBIT (0x8000000000000000L)
#define BN_DEC_CONV (10000000000000000000UL)
#define BN_DEC_FMT1 "%lu"
#define BN_DEC_FMT2 "%019lu"
#define BN_DEC_NUM 19
#define BN_HEX_FMT1 "%lX"
#define BN_HEX_FMT2 "%016lX"
#endif
/* This is where the long long data type is 64 bits, but long is 32.
* For machines where there are 64bit registers, this is the mode to use.
* IRIX, on R4000 and above should use this mode, along with the relevant
* assembler code :-). Do NOT define BN_LLONG.
*/
#ifdef SIXTY_FOUR_BIT
#undef BN_LLONG
#undef BN_ULLONG
#define BN_ULONG unsigned long long
#define BN_LONG long long
#define BN_BITS 128
#define BN_BYTES 8
#define BN_BITS2 64
#define BN_BITS4 32
#define BN_MASK2 (0xffffffffffffffffLL)
#define BN_MASK2l (0xffffffffL)
#define BN_MASK2h (0xffffffff00000000LL)
#define BN_MASK2h1 (0xffffffff80000000LL)
#define BN_TBIT (0x8000000000000000LL)
#define BN_DEC_CONV (10000000000000000000ULL)
#define BN_DEC_FMT1 "%llu"
#define BN_DEC_FMT2 "%019llu"
#define BN_DEC_NUM 19
#define BN_HEX_FMT1 "%llX"
#define BN_HEX_FMT2 "%016llX"
#endif
#ifdef THIRTY_TWO_BIT
#ifdef BN_LLONG
# if defined(_WIN32) && !defined(__GNUC__)
# define BN_ULLONG unsigned __int64
# define BN_MASK (0xffffffffffffffffI64)
# else
# define BN_ULLONG unsigned long long
# define BN_MASK (0xffffffffffffffffLL)
# endif
#endif
#define BN_ULONG unsigned int
#define BN_LONG int
#define BN_BITS 64
#define BN_BYTES 4
#define BN_BITS2 32
#define BN_BITS4 16
#define BN_MASK2 (0xffffffffL)
#define BN_MASK2l (0xffff)
#define BN_MASK2h1 (0xffff8000L)
#define BN_MASK2h (0xffff0000L)
#define BN_TBIT (0x80000000L)
#define BN_DEC_CONV (1000000000L)
#define BN_DEC_FMT1 "%u"
#define BN_DEC_FMT2 "%09u"
#define BN_DEC_NUM 9
#define BN_HEX_FMT1 "%X"
#define BN_HEX_FMT2 "%08X"
#endif
/* 2011-02-22 SMS.
* In various places, a size_t variable or a type cast to size_t was
* used to perform integer-only operations on pointers. This failed on
* VMS with 64-bit pointers (CC /POINTER_SIZE = 64) because size_t is
* still only 32 bits. What's needed in these cases is an integer type
* with the same size as a pointer, which size_t is not certain to be.
* The only fix here is VMS-specific.
*/
#if defined(OPENSSL_SYS_VMS)
# if __INITIAL_POINTER_SIZE == 64
# define PTR_SIZE_INT long long
# else /* __INITIAL_POINTER_SIZE == 64 */
# define PTR_SIZE_INT int
# endif /* __INITIAL_POINTER_SIZE == 64 [else] */
#else /* defined(OPENSSL_SYS_VMS) */
# define PTR_SIZE_INT size_t
#endif /* defined(OPENSSL_SYS_VMS) [else] */
#define BN_DEFAULT_BITS 1280
#define BN_FLG_MALLOCED 0x01
#define BN_FLG_STATIC_DATA 0x02
#define BN_FLG_CONSTTIME 0x04 /* avoid leaking exponent information through timing,
* BN_mod_exp_mont() will call BN_mod_exp_mont_consttime,
* BN_div() will call BN_div_no_branch,
* BN_mod_inverse() will call BN_mod_inverse_no_branch.
*/
#ifndef OPENSSL_NO_DEPRECATED
#define BN_FLG_EXP_CONSTTIME BN_FLG_CONSTTIME /* deprecated name for the flag */
/* avoid leaking exponent information through timings
* (BN_mod_exp_mont() will call BN_mod_exp_mont_consttime) */
#endif
#ifndef OPENSSL_NO_DEPRECATED
#define BN_FLG_FREE 0x8000 /* used for debuging */
#endif
#define BN_set_flags(b,n) ((b)->flags|=(n))
#define BN_get_flags(b,n) ((b)->flags&(n))
/* get a clone of a BIGNUM with changed flags, for *temporary* use only
* (the two BIGNUMs cannot not be used in parallel!) */
#define BN_with_flags(dest,b,n) ((dest)->d=(b)->d, \
(dest)->top=(b)->top, \
(dest)->dmax=(b)->dmax, \
(dest)->neg=(b)->neg, \
(dest)->flags=(((dest)->flags & BN_FLG_MALLOCED) \
| ((b)->flags & ~BN_FLG_MALLOCED) \
| BN_FLG_STATIC_DATA \
| (n)))
/* Already declared in ossl_typ.h */
#if 0
typedef struct bignum_st BIGNUM;
/* Used for temp variables (declaration hidden in bn_lcl.h) */
typedef struct bignum_ctx BN_CTX;
typedef struct bn_blinding_st BN_BLINDING;
typedef struct bn_mont_ctx_st BN_MONT_CTX;
typedef struct bn_recp_ctx_st BN_RECP_CTX;
typedef struct bn_gencb_st BN_GENCB;
#endif
struct bignum_st
{
BN_ULONG *d; /* Pointer to an array of 'BN_BITS2' bit chunks. */
int top; /* Index of last used d +1. */
/* The next are internal book keeping for bn_expand. */
int dmax; /* Size of the d array. */
int neg; /* one if the number is negative */
int flags;
};
/* Used for montgomery multiplication */
struct bn_mont_ctx_st
{
int ri; /* number of bits in R */
BIGNUM RR; /* used to convert to montgomery form */
BIGNUM N; /* The modulus */
BIGNUM Ni; /* R*(1/R mod N) - N*Ni = 1
* (Ni is only stored for bignum algorithm) */
BN_ULONG n0[2];/* least significant word(s) of Ni;
(type changed with 0.9.9, was "BN_ULONG n0;" before) */
int flags;
};
/* Used for reciprocal division/mod functions
* It cannot be shared between threads
*/
struct bn_recp_ctx_st
{
BIGNUM N; /* the divisor */
BIGNUM Nr; /* the reciprocal */
int num_bits;
int shift;
int flags;
};
/* Used for slow "generation" functions. */
struct bn_gencb_st
{
unsigned int ver; /* To handle binary (in)compatibility */
void *arg; /* callback-specific data */
union
{
/* if(ver==1) - handles old style callbacks */
void (*cb_1)(int, int, void *);
/* if(ver==2) - new callback style */
int (*cb_2)(int, int, BN_GENCB *);
} cb;
};
/* Wrapper function to make using BN_GENCB easier, */
int BN_GENCB_call(BN_GENCB *cb, int a, int b);
/* Macro to populate a BN_GENCB structure with an "old"-style callback */
#define BN_GENCB_set_old(gencb, callback, cb_arg) { \
BN_GENCB *tmp_gencb = (gencb); \
tmp_gencb->ver = 1; \
tmp_gencb->arg = (cb_arg); \
tmp_gencb->cb.cb_1 = (callback); }
/* Macro to populate a BN_GENCB structure with a "new"-style callback */
#define BN_GENCB_set(gencb, callback, cb_arg) { \
BN_GENCB *tmp_gencb = (gencb); \
tmp_gencb->ver = 2; \
tmp_gencb->arg = (cb_arg); \
tmp_gencb->cb.cb_2 = (callback); }
#define BN_prime_checks 0 /* default: select number of iterations
based on the size of the number */
/* number of Miller-Rabin iterations for an error rate of less than 2^-80
* for random 'b'-bit input, b >= 100 (taken from table 4.4 in the Handbook
* of Applied Cryptography [Menezes, van Oorschot, Vanstone; CRC Press 1996];
* original paper: Damgaard, Landrock, Pomerance: Average case error estimates
* for the strong probable prime test. -- Math. Comp. 61 (1993) 177-194) */
#define BN_prime_checks_for_size(b) ((b) >= 1300 ? 2 : \
(b) >= 850 ? 3 : \
(b) >= 650 ? 4 : \
(b) >= 550 ? 5 : \
(b) >= 450 ? 6 : \
(b) >= 400 ? 7 : \
(b) >= 350 ? 8 : \
(b) >= 300 ? 9 : \
(b) >= 250 ? 12 : \
(b) >= 200 ? 15 : \
(b) >= 150 ? 18 : \
/* b >= 100 */ 27)
#define BN_num_bytes(a) ((BN_num_bits(a)+7)/8)
/* Note that BN_abs_is_word didn't work reliably for w == 0 until 0.9.8 */
#define BN_abs_is_word(a,w) ((((a)->top == 1) && ((a)->d[0] == (BN_ULONG)(w))) || \
(((w) == 0) && ((a)->top == 0)))
#define BN_is_zero(a) ((a)->top == 0)
#define BN_is_one(a) (BN_abs_is_word((a),1) && !(a)->neg)
#define BN_is_word(a,w) (BN_abs_is_word((a),(w)) && (!(w) || !(a)->neg))
#define BN_is_odd(a) (((a)->top > 0) && ((a)->d[0] & 1))
#define BN_one(a) (BN_set_word((a),1))
#define BN_zero_ex(a) \
do { \
BIGNUM *_tmp_bn = (a); \
_tmp_bn->top = 0; \
_tmp_bn->neg = 0; \
} while(0)
#ifdef OPENSSL_NO_DEPRECATED
#define BN_zero(a) BN_zero_ex(a)
#else
#define BN_zero(a) (BN_set_word((a),0))
#endif
const BIGNUM *BN_value_one(void);
char * BN_options(void);
BN_CTX *BN_CTX_new(void);
#ifndef OPENSSL_NO_DEPRECATED
void BN_CTX_init(BN_CTX *c);
#endif
void BN_CTX_free(BN_CTX *c);
void BN_CTX_start(BN_CTX *ctx);
BIGNUM *BN_CTX_get(BN_CTX *ctx);
void BN_CTX_end(BN_CTX *ctx);
int BN_rand(BIGNUM *rnd, int bits, int top,int bottom);
int BN_pseudo_rand(BIGNUM *rnd, int bits, int top,int bottom);
int BN_rand_range(BIGNUM *rnd, const BIGNUM *range);
int BN_pseudo_rand_range(BIGNUM *rnd, const BIGNUM *range);
int BN_num_bits(const BIGNUM *a);
int BN_num_bits_word(BN_ULONG);
BIGNUM *BN_new(void);
void BN_init(BIGNUM *);
void BN_clear_free(BIGNUM *a);
BIGNUM *BN_copy(BIGNUM *a, const BIGNUM *b);
void BN_swap(BIGNUM *a, BIGNUM *b);
BIGNUM *BN_bin2bn(const unsigned char *s,int len,BIGNUM *ret);
int BN_bn2bin(const BIGNUM *a, unsigned char *to);
BIGNUM *BN_mpi2bn(const unsigned char *s,int len,BIGNUM *ret);
int BN_bn2mpi(const BIGNUM *a, unsigned char *to);
int BN_sub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b);
int BN_usub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b);
int BN_uadd(BIGNUM *r, const BIGNUM *a, const BIGNUM *b);
int BN_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b);
int BN_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, BN_CTX *ctx);
int BN_sqr(BIGNUM *r, const BIGNUM *a,BN_CTX *ctx);
/** BN_set_negative sets sign of a BIGNUM
* \param b pointer to the BIGNUM object
* \param n 0 if the BIGNUM b should be positive and a value != 0 otherwise
*/
void BN_set_negative(BIGNUM *b, int n);
/** BN_is_negative returns 1 if the BIGNUM is negative
* \param a pointer to the BIGNUM object
* \return 1 if a < 0 and 0 otherwise
*/
#define BN_is_negative(a) ((a)->neg != 0)
int BN_div(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m, const BIGNUM *d,
BN_CTX *ctx);
#define BN_mod(rem,m,d,ctx) BN_div(NULL,(rem),(m),(d),(ctx))
int BN_nnmod(BIGNUM *r, const BIGNUM *m, const BIGNUM *d, BN_CTX *ctx);
int BN_mod_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m, BN_CTX *ctx);
int BN_mod_add_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m);
int BN_mod_sub(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m, BN_CTX *ctx);
int BN_mod_sub_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *b, const BIGNUM *m);
int BN_mod_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const BIGNUM *m, BN_CTX *ctx);
int BN_mod_sqr(BIGNUM *r, const BIGNUM *a, const BIGNUM *m, BN_CTX *ctx);
int BN_mod_lshift1(BIGNUM *r, const BIGNUM *a, const BIGNUM *m, BN_CTX *ctx);
int BN_mod_lshift1_quick(BIGNUM *r, const BIGNUM *a, const BIGNUM *m);
int BN_mod_lshift(BIGNUM *r, const BIGNUM *a, int n, const BIGNUM *m, BN_CTX *ctx);
int BN_mod_lshift_quick(BIGNUM *r, const BIGNUM *a, int n, const BIGNUM *m);
BN_ULONG BN_mod_word(const BIGNUM *a, BN_ULONG w);
BN_ULONG BN_div_word(BIGNUM *a, BN_ULONG w);
int BN_mul_word(BIGNUM *a, BN_ULONG w);
int BN_add_word(BIGNUM *a, BN_ULONG w);
int BN_sub_word(BIGNUM *a, BN_ULONG w);
int BN_set_word(BIGNUM *a, BN_ULONG w);
BN_ULONG BN_get_word(const BIGNUM *a);
int BN_cmp(const BIGNUM *a, const BIGNUM *b);
void BN_free(BIGNUM *a);
int BN_is_bit_set(const BIGNUM *a, int n);
int BN_lshift(BIGNUM *r, const BIGNUM *a, int n);
int BN_lshift1(BIGNUM *r, const BIGNUM *a);
int BN_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,BN_CTX *ctx);
int BN_mod_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m,BN_CTX *ctx);
int BN_mod_exp_mont(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *m_ctx);
int BN_mod_exp_mont_consttime(BIGNUM *rr, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *in_mont);
int BN_mod_exp_mont_word(BIGNUM *r, BN_ULONG a, const BIGNUM *p,
const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *m_ctx);
int BN_mod_exp2_mont(BIGNUM *r, const BIGNUM *a1, const BIGNUM *p1,
const BIGNUM *a2, const BIGNUM *p2,const BIGNUM *m,
BN_CTX *ctx,BN_MONT_CTX *m_ctx);
int BN_mod_exp_simple(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m,BN_CTX *ctx);
int BN_mask_bits(BIGNUM *a,int n);
#ifndef OPENSSL_NO_FP_API
int BN_print_fp(FILE *fp, const BIGNUM *a);
#endif
#ifdef HEADER_BIO_H
int BN_print(BIO *fp, const BIGNUM *a);
#else
int BN_print(void *fp, const BIGNUM *a);
#endif
int BN_reciprocal(BIGNUM *r, const BIGNUM *m, int len, BN_CTX *ctx);
int BN_rshift(BIGNUM *r, const BIGNUM *a, int n);
int BN_rshift1(BIGNUM *r, const BIGNUM *a);
void BN_clear(BIGNUM *a);
BIGNUM *BN_dup(const BIGNUM *a);
int BN_ucmp(const BIGNUM *a, const BIGNUM *b);
int BN_set_bit(BIGNUM *a, int n);
int BN_clear_bit(BIGNUM *a, int n);
char * BN_bn2hex(const BIGNUM *a);
char * BN_bn2dec(const BIGNUM *a);
int BN_hex2bn(BIGNUM **a, const char *str);
int BN_dec2bn(BIGNUM **a, const char *str);
int BN_asc2bn(BIGNUM **a, const char *str);
int BN_gcd(BIGNUM *r,const BIGNUM *a,const BIGNUM *b,BN_CTX *ctx);
int BN_kronecker(const BIGNUM *a,const BIGNUM *b,BN_CTX *ctx); /* returns -2 for error */
BIGNUM *BN_mod_inverse(BIGNUM *ret,
const BIGNUM *a, const BIGNUM *n,BN_CTX *ctx);
BIGNUM *BN_mod_sqrt(BIGNUM *ret,
const BIGNUM *a, const BIGNUM *n,BN_CTX *ctx);
void BN_consttime_swap(BN_ULONG swap, BIGNUM *a, BIGNUM *b, int nwords);
/* Deprecated versions */
#ifndef OPENSSL_NO_DEPRECATED
BIGNUM *BN_generate_prime(BIGNUM *ret,int bits,int safe,
const BIGNUM *add, const BIGNUM *rem,
void (*callback)(int,int,void *),void *cb_arg);
int BN_is_prime(const BIGNUM *p,int nchecks,
void (*callback)(int,int,void *),
BN_CTX *ctx,void *cb_arg);
int BN_is_prime_fasttest(const BIGNUM *p,int nchecks,
void (*callback)(int,int,void *),BN_CTX *ctx,void *cb_arg,
int do_trial_division);
#endif /* !defined(OPENSSL_NO_DEPRECATED) */
/* Newer versions */
int BN_generate_prime_ex(BIGNUM *ret,int bits,int safe, const BIGNUM *add,
const BIGNUM *rem, BN_GENCB *cb);
int BN_is_prime_ex(const BIGNUM *p,int nchecks, BN_CTX *ctx, BN_GENCB *cb);
int BN_is_prime_fasttest_ex(const BIGNUM *p,int nchecks, BN_CTX *ctx,
int do_trial_division, BN_GENCB *cb);
int BN_X931_generate_Xpq(BIGNUM *Xp, BIGNUM *Xq, int nbits, BN_CTX *ctx);
int BN_X931_derive_prime_ex(BIGNUM *p, BIGNUM *p1, BIGNUM *p2,
const BIGNUM *Xp, const BIGNUM *Xp1, const BIGNUM *Xp2,
const BIGNUM *e, BN_CTX *ctx, BN_GENCB *cb);
int BN_X931_generate_prime_ex(BIGNUM *p, BIGNUM *p1, BIGNUM *p2,
BIGNUM *Xp1, BIGNUM *Xp2,
const BIGNUM *Xp,
const BIGNUM *e, BN_CTX *ctx,
BN_GENCB *cb);
BN_MONT_CTX *BN_MONT_CTX_new(void );
void BN_MONT_CTX_init(BN_MONT_CTX *ctx);
int BN_mod_mul_montgomery(BIGNUM *r,const BIGNUM *a,const BIGNUM *b,
BN_MONT_CTX *mont, BN_CTX *ctx);
#define BN_to_montgomery(r,a,mont,ctx) BN_mod_mul_montgomery(\
(r),(a),&((mont)->RR),(mont),(ctx))
int BN_from_montgomery(BIGNUM *r,const BIGNUM *a,
BN_MONT_CTX *mont, BN_CTX *ctx);
void BN_MONT_CTX_free(BN_MONT_CTX *mont);
int BN_MONT_CTX_set(BN_MONT_CTX *mont,const BIGNUM *mod,BN_CTX *ctx);
BN_MONT_CTX *BN_MONT_CTX_copy(BN_MONT_CTX *to,BN_MONT_CTX *from);
BN_MONT_CTX *BN_MONT_CTX_set_locked(BN_MONT_CTX **pmont, int lock,
const BIGNUM *mod, BN_CTX *ctx);
/* BN_BLINDING flags */
#define BN_BLINDING_NO_UPDATE 0x00000001
#define BN_BLINDING_NO_RECREATE 0x00000002
BN_BLINDING *BN_BLINDING_new(const BIGNUM *A, const BIGNUM *Ai, BIGNUM *mod);
void BN_BLINDING_free(BN_BLINDING *b);
int BN_BLINDING_update(BN_BLINDING *b,BN_CTX *ctx);
int BN_BLINDING_convert(BIGNUM *n, BN_BLINDING *b, BN_CTX *ctx);
int BN_BLINDING_invert(BIGNUM *n, BN_BLINDING *b, BN_CTX *ctx);
int BN_BLINDING_convert_ex(BIGNUM *n, BIGNUM *r, BN_BLINDING *b, BN_CTX *);
int BN_BLINDING_invert_ex(BIGNUM *n, const BIGNUM *r, BN_BLINDING *b, BN_CTX *);
#ifndef OPENSSL_NO_DEPRECATED
unsigned long BN_BLINDING_get_thread_id(const BN_BLINDING *);
void BN_BLINDING_set_thread_id(BN_BLINDING *, unsigned long);
#endif
CRYPTO_THREADID *BN_BLINDING_thread_id(BN_BLINDING *);
unsigned long BN_BLINDING_get_flags(const BN_BLINDING *);
void BN_BLINDING_set_flags(BN_BLINDING *, unsigned long);
BN_BLINDING *BN_BLINDING_create_param(BN_BLINDING *b,
const BIGNUM *e, BIGNUM *m, BN_CTX *ctx,
int (*bn_mod_exp)(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m, BN_CTX *ctx, BN_MONT_CTX *m_ctx),
BN_MONT_CTX *m_ctx);
#ifndef OPENSSL_NO_DEPRECATED
void BN_set_params(int mul,int high,int low,int mont);
int BN_get_params(int which); /* 0, mul, 1 high, 2 low, 3 mont */
#endif
void BN_RECP_CTX_init(BN_RECP_CTX *recp);
BN_RECP_CTX *BN_RECP_CTX_new(void);
void BN_RECP_CTX_free(BN_RECP_CTX *recp);
int BN_RECP_CTX_set(BN_RECP_CTX *recp,const BIGNUM *rdiv,BN_CTX *ctx);
int BN_mod_mul_reciprocal(BIGNUM *r, const BIGNUM *x, const BIGNUM *y,
BN_RECP_CTX *recp,BN_CTX *ctx);
int BN_mod_exp_recp(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
const BIGNUM *m, BN_CTX *ctx);
int BN_div_recp(BIGNUM *dv, BIGNUM *rem, const BIGNUM *m,
BN_RECP_CTX *recp, BN_CTX *ctx);
#ifndef OPENSSL_NO_EC2M
/* Functions for arithmetic over binary polynomials represented by BIGNUMs.
*
* The BIGNUM::neg property of BIGNUMs representing binary polynomials is
* ignored.
*
* Note that input arguments are not const so that their bit arrays can
* be expanded to the appropriate size if needed.
*/
int BN_GF2m_add(BIGNUM *r, const BIGNUM *a, const BIGNUM *b); /*r = a + b*/
#define BN_GF2m_sub(r, a, b) BN_GF2m_add(r, a, b)
int BN_GF2m_mod(BIGNUM *r, const BIGNUM *a, const BIGNUM *p); /*r=a mod p*/
int BN_GF2m_mod_mul(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const BIGNUM *p, BN_CTX *ctx); /* r = (a * b) mod p */
int BN_GF2m_mod_sqr(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
BN_CTX *ctx); /* r = (a * a) mod p */
int BN_GF2m_mod_inv(BIGNUM *r, const BIGNUM *b, const BIGNUM *p,
BN_CTX *ctx); /* r = (1 / b) mod p */
int BN_GF2m_mod_div(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const BIGNUM *p, BN_CTX *ctx); /* r = (a / b) mod p */
int BN_GF2m_mod_exp(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const BIGNUM *p, BN_CTX *ctx); /* r = (a ^ b) mod p */
int BN_GF2m_mod_sqrt(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
BN_CTX *ctx); /* r = sqrt(a) mod p */
int BN_GF2m_mod_solve_quad(BIGNUM *r, const BIGNUM *a, const BIGNUM *p,
BN_CTX *ctx); /* r^2 + r = a mod p */
#define BN_GF2m_cmp(a, b) BN_ucmp((a), (b))
/* Some functions allow for representation of the irreducible polynomials
* as an unsigned int[], say p. The irreducible f(t) is then of the form:
* t^p[0] + t^p[1] + ... + t^p[k]
* where m = p[0] > p[1] > ... > p[k] = 0.
*/
int BN_GF2m_mod_arr(BIGNUM *r, const BIGNUM *a, const int p[]);
/* r = a mod p */
int BN_GF2m_mod_mul_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const int p[], BN_CTX *ctx); /* r = (a * b) mod p */
int BN_GF2m_mod_sqr_arr(BIGNUM *r, const BIGNUM *a, const int p[],
BN_CTX *ctx); /* r = (a * a) mod p */
int BN_GF2m_mod_inv_arr(BIGNUM *r, const BIGNUM *b, const int p[],
BN_CTX *ctx); /* r = (1 / b) mod p */
int BN_GF2m_mod_div_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const int p[], BN_CTX *ctx); /* r = (a / b) mod p */
int BN_GF2m_mod_exp_arr(BIGNUM *r, const BIGNUM *a, const BIGNUM *b,
const int p[], BN_CTX *ctx); /* r = (a ^ b) mod p */
int BN_GF2m_mod_sqrt_arr(BIGNUM *r, const BIGNUM *a,
const int p[], BN_CTX *ctx); /* r = sqrt(a) mod p */
int BN_GF2m_mod_solve_quad_arr(BIGNUM *r, const BIGNUM *a,
const int p[], BN_CTX *ctx); /* r^2 + r = a mod p */
int BN_GF2m_poly2arr(const BIGNUM *a, int p[], int max);
int BN_GF2m_arr2poly(const int p[], BIGNUM *a);
#endif
/* faster mod functions for the 'NIST primes'
* 0 <= a < p^2 */
int BN_nist_mod_192(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx);
int BN_nist_mod_224(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx);
int BN_nist_mod_256(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx);
int BN_nist_mod_384(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx);
int BN_nist_mod_521(BIGNUM *r, const BIGNUM *a, const BIGNUM *p, BN_CTX *ctx);
const BIGNUM *BN_get0_nist_prime_192(void);
const BIGNUM *BN_get0_nist_prime_224(void);
const BIGNUM *BN_get0_nist_prime_256(void);
const BIGNUM *BN_get0_nist_prime_384(void);
const BIGNUM *BN_get0_nist_prime_521(void);
/* library internal functions */
#define bn_expand(a,bits) ((((((bits+BN_BITS2-1))/BN_BITS2)) <= (a)->dmax)?\
(a):bn_expand2((a),(bits+BN_BITS2-1)/BN_BITS2))
#define bn_wexpand(a,words) (((words) <= (a)->dmax)?(a):bn_expand2((a),(words)))
BIGNUM *bn_expand2(BIGNUM *a, int words);
#ifndef OPENSSL_NO_DEPRECATED
BIGNUM *bn_dup_expand(const BIGNUM *a, int words); /* unused */
#endif
/* Bignum consistency macros
* There is one "API" macro, bn_fix_top(), for stripping leading zeroes from
* bignum data after direct manipulations on the data. There is also an
* "internal" macro, bn_check_top(), for verifying that there are no leading
* zeroes. Unfortunately, some auditing is required due to the fact that
* bn_fix_top() has become an overabused duct-tape because bignum data is
* occasionally passed around in an inconsistent state. So the following
* changes have been made to sort this out;
* - bn_fix_top()s implementation has been moved to bn_correct_top()
* - if BN_DEBUG isn't defined, bn_fix_top() maps to bn_correct_top(), and
* bn_check_top() is as before.
* - if BN_DEBUG *is* defined;
* - bn_check_top() tries to pollute unused words even if the bignum 'top' is
* consistent. (ed: only if BN_DEBUG_RAND is defined)
* - bn_fix_top() maps to bn_check_top() rather than "fixing" anything.
* The idea is to have debug builds flag up inconsistent bignums when they
* occur. If that occurs in a bn_fix_top(), we examine the code in question; if
* the use of bn_fix_top() was appropriate (ie. it follows directly after code
* that manipulates the bignum) it is converted to bn_correct_top(), and if it
* was not appropriate, we convert it permanently to bn_check_top() and track
* down the cause of the bug. Eventually, no internal code should be using the
* bn_fix_top() macro. External applications and libraries should try this with
* their own code too, both in terms of building against the openssl headers
* with BN_DEBUG defined *and* linking with a version of OpenSSL built with it
* defined. This not only improves external code, it provides more test
* coverage for openssl's own code.
*/
#ifdef BN_DEBUG
/* We only need assert() when debugging */
#include <assert.h>
#ifdef BN_DEBUG_RAND
/* To avoid "make update" cvs wars due to BN_DEBUG, use some tricks */
#ifndef RAND_pseudo_bytes
int RAND_pseudo_bytes(unsigned char *buf,int num);
#define BN_DEBUG_TRIX
#endif
#define bn_pollute(a) \
do { \
const BIGNUM *_bnum1 = (a); \
if(_bnum1->top < _bnum1->dmax) { \
unsigned char _tmp_char; \
/* We cast away const without the compiler knowing, any \
* *genuinely* constant variables that aren't mutable \
* wouldn't be constructed with top!=dmax. */ \
BN_ULONG *_not_const; \
memcpy(&_not_const, &_bnum1->d, sizeof(BN_ULONG*)); \
RAND_pseudo_bytes(&_tmp_char, 1); \
memset((unsigned char *)(_not_const + _bnum1->top), _tmp_char, \
(_bnum1->dmax - _bnum1->top) * sizeof(BN_ULONG)); \
} \
} while(0)
#ifdef BN_DEBUG_TRIX
#undef RAND_pseudo_bytes
#endif
#else
#define bn_pollute(a)
#endif
#define bn_check_top(a) \
do { \
const BIGNUM *_bnum2 = (a); \
if (_bnum2 != NULL) { \
assert((_bnum2->top == 0) || \
(_bnum2->d[_bnum2->top - 1] != 0)); \
bn_pollute(_bnum2); \
} \
} while(0)
#define bn_fix_top(a) bn_check_top(a)
#define bn_check_size(bn, bits) bn_wcheck_size(bn, ((bits+BN_BITS2-1))/BN_BITS2)
#define bn_wcheck_size(bn, words) \
do { \
const BIGNUM *_bnum2 = (bn); \
assert(words <= (_bnum2)->dmax && words >= (_bnum2)->top); \
} while(0)
#else /* !BN_DEBUG */
#define bn_pollute(a)
#define bn_check_top(a)
#define bn_fix_top(a) bn_correct_top(a)
#define bn_check_size(bn, bits)
#define bn_wcheck_size(bn, words)
#endif
#define bn_correct_top(a) \
{ \
BN_ULONG *ftl; \
int tmp_top = (a)->top; \
if (tmp_top > 0) \
{ \
for (ftl= &((a)->d[tmp_top-1]); tmp_top > 0; tmp_top--) \
if (*(ftl--)) break; \
(a)->top = tmp_top; \
} \
bn_pollute(a); \
}
BN_ULONG bn_mul_add_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w);
BN_ULONG bn_mul_words(BN_ULONG *rp, const BN_ULONG *ap, int num, BN_ULONG w);
void bn_sqr_words(BN_ULONG *rp, const BN_ULONG *ap, int num);
BN_ULONG bn_div_words(BN_ULONG h, BN_ULONG l, BN_ULONG d);
BN_ULONG bn_add_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int num);
BN_ULONG bn_sub_words(BN_ULONG *rp, const BN_ULONG *ap, const BN_ULONG *bp,int num);
/* Primes from RFC 2409 */
BIGNUM *get_rfc2409_prime_768(BIGNUM *bn);
BIGNUM *get_rfc2409_prime_1024(BIGNUM *bn);
/* Primes from RFC 3526 */
BIGNUM *get_rfc3526_prime_1536(BIGNUM *bn);
BIGNUM *get_rfc3526_prime_2048(BIGNUM *bn);
BIGNUM *get_rfc3526_prime_3072(BIGNUM *bn);
BIGNUM *get_rfc3526_prime_4096(BIGNUM *bn);
BIGNUM *get_rfc3526_prime_6144(BIGNUM *bn);
BIGNUM *get_rfc3526_prime_8192(BIGNUM *bn);
int BN_bntest_rand(BIGNUM *rnd, int bits, int top,int bottom);
/* BEGIN ERROR CODES */
/* The following lines are auto generated by the script mkerr.pl. Any changes
* made after this point may be overwritten when the script is next run.
*/
void ERR_load_BN_strings(void);
/* Error codes for the BN functions. */
/* Function codes. */
#define BN_F_BNRAND 127
#define BN_F_BN_BLINDING_CONVERT_EX 100
#define BN_F_BN_BLINDING_CREATE_PARAM 128
#define BN_F_BN_BLINDING_INVERT_EX 101
#define BN_F_BN_BLINDING_NEW 102
#define BN_F_BN_BLINDING_UPDATE 103
#define BN_F_BN_BN2DEC 104
#define BN_F_BN_BN2HEX 105
#define BN_F_BN_CTX_GET 116
#define BN_F_BN_CTX_NEW 106
#define BN_F_BN_CTX_START 129
#define BN_F_BN_DIV 107
#define BN_F_BN_DIV_NO_BRANCH 138
#define BN_F_BN_DIV_RECP 130
#define BN_F_BN_EXP 123
#define BN_F_BN_EXPAND2 108
#define BN_F_BN_EXPAND_INTERNAL 120
#define BN_F_BN_GF2M_MOD 131
#define BN_F_BN_GF2M_MOD_EXP 132
#define BN_F_BN_GF2M_MOD_MUL 133
#define BN_F_BN_GF2M_MOD_SOLVE_QUAD 134
#define BN_F_BN_GF2M_MOD_SOLVE_QUAD_ARR 135
#define BN_F_BN_GF2M_MOD_SQR 136
#define BN_F_BN_GF2M_MOD_SQRT 137
#define BN_F_BN_MOD_EXP2_MONT 118
#define BN_F_BN_MOD_EXP_MONT 109
#define BN_F_BN_MOD_EXP_MONT_CONSTTIME 124
#define BN_F_BN_MOD_EXP_MONT_WORD 117
#define BN_F_BN_MOD_EXP_RECP 125
#define BN_F_BN_MOD_EXP_SIMPLE 126
#define BN_F_BN_MOD_INVERSE 110
#define BN_F_BN_MOD_INVERSE_NO_BRANCH 139
#define BN_F_BN_MOD_LSHIFT_QUICK 119
#define BN_F_BN_MOD_MUL_RECIPROCAL 111
#define BN_F_BN_MOD_SQRT 121
#define BN_F_BN_MPI2BN 112
#define BN_F_BN_NEW 113
#define BN_F_BN_RAND 114
#define BN_F_BN_RAND_RANGE 122
#define BN_F_BN_USUB 115
/* Reason codes. */
#define BN_R_ARG2_LT_ARG3 100
#define BN_R_BAD_RECIPROCAL 101
#define BN_R_BIGNUM_TOO_LONG 114
#define BN_R_CALLED_WITH_EVEN_MODULUS 102
#define BN_R_DIV_BY_ZERO 103
#define BN_R_ENCODING_ERROR 104
#define BN_R_EXPAND_ON_STATIC_BIGNUM_DATA 105
#define BN_R_INPUT_NOT_REDUCED 110
#define BN_R_INVALID_LENGTH 106
#define BN_R_INVALID_RANGE 115
#define BN_R_NOT_A_SQUARE 111
#define BN_R_NOT_INITIALIZED 107
#define BN_R_NO_INVERSE 108
#define BN_R_NO_SOLUTION 116
#define BN_R_P_IS_NOT_PRIME 112
#define BN_R_TOO_MANY_ITERATIONS 113
#define BN_R_TOO_MANY_TEMPORARY_VARIABLES 109
#ifdef __cplusplus
}
#endif
#endif
@@ -1,611 +0,0 @@
/* crypto/crypto.h */
/* ====================================================================
* Copyright (c) 1998-2006 The OpenSSL Project. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* 3. All advertising materials mentioning features or use of this
* software must display the following acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
*
* 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
* endorse or promote products derived from this software without
* prior written permission. For written permission, please contact
* openssl-core@openssl.org.
*
* 5. Products derived from this software may not be called "OpenSSL"
* nor may "OpenSSL" appear in their names without prior written
* permission of the OpenSSL Project.
*
* 6. Redistributions of any form whatsoever must retain the following
* acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit (http://www.openssl.org/)"
*
* THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
* EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
* ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
* OF THE POSSIBILITY OF SUCH DAMAGE.
* ====================================================================
*
* This product includes cryptographic software written by Eric Young
* (eay@cryptsoft.com). This product includes software written by Tim
* Hudson (tjh@cryptsoft.com).
*
*/
/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
* All rights reserved.
*
* This package is an SSL implementation written
* by Eric Young (eay@cryptsoft.com).
* The implementation was written so as to conform with Netscapes SSL.
*
* This library is free for commercial and non-commercial use as long as
* the following conditions are aheared to. The following conditions
* apply to all code found in this distribution, be it the RC4, RSA,
* lhash, DES, etc., code; not just the SSL code. The SSL documentation
* included with this distribution is covered by the same copyright terms
* except that the holder is Tim Hudson (tjh@cryptsoft.com).
*
* Copyright remains Eric Young's, and as such any Copyright notices in
* the code are not to be removed.
* If this package is used in a product, Eric Young should be given attribution
* as the author of the parts of the library used.
* This can be in the form of a textual message at program startup or
* in documentation (online or textual) provided with the package.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* "This product includes cryptographic software written by
* Eric Young (eay@cryptsoft.com)"
* The word 'cryptographic' can be left out if the rouines from the library
* being used are not cryptographic related :-).
* 4. If you include any Windows specific code (or a derivative thereof) from
* the apps directory (application code) you must include an acknowledgement:
* "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
*
* THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* The licence and distribution terms for any publically available version or
* derivative of this code cannot be changed. i.e. this code cannot simply be
* copied and put under another distribution licence
* [including the GNU Public Licence.]
*/
/* ====================================================================
* Copyright 2002 Sun Microsystems, Inc. ALL RIGHTS RESERVED.
* ECDH support in OpenSSL originally developed by
* SUN MICROSYSTEMS, INC., and contributed to the OpenSSL project.
*/
#ifndef HEADER_CRYPTO_H
#define HEADER_CRYPTO_H
#include <stdlib.h>
#include <openssl/e_os2.h>
#ifndef OPENSSL_NO_FP_API
#include <stdio.h>
#endif
#include <openssl/stack.h>
#include <openssl/safestack.h>
#include <openssl/opensslv.h>
#include <openssl/ossl_typ.h>
#ifdef CHARSET_EBCDIC
#include <openssl/ebcdic.h>
#endif
/* Resolve problems on some operating systems with symbol names that clash
one way or another */
#include <openssl/symhacks.h>
#ifdef __cplusplus
extern "C" {
#endif
/* Backward compatibility to SSLeay */
/* This is more to be used to check the correct DLL is being used
* in the MS world. */
#define SSLEAY_VERSION_NUMBER OPENSSL_VERSION_NUMBER
#define SSLEAY_VERSION 0
/* #define SSLEAY_OPTIONS 1 no longer supported */
#define SSLEAY_CFLAGS 2
#define SSLEAY_BUILT_ON 3
#define SSLEAY_PLATFORM 4
#define SSLEAY_DIR 5
/* Already declared in ossl_typ.h */
#if 0
typedef struct crypto_ex_data_st CRYPTO_EX_DATA;
/* Called when a new object is created */
typedef int CRYPTO_EX_new(void *parent, void *ptr, CRYPTO_EX_DATA *ad,
int idx, long argl, void *argp);
/* Called when an object is free()ed */
typedef void CRYPTO_EX_free(void *parent, void *ptr, CRYPTO_EX_DATA *ad,
int idx, long argl, void *argp);
/* Called when we need to dup an object */
typedef int CRYPTO_EX_dup(CRYPTO_EX_DATA *to, CRYPTO_EX_DATA *from, void *from_d,
int idx, long argl, void *argp);
#endif
/* A generic structure to pass assorted data in a expandable way */
typedef struct openssl_item_st
{
int code;
void *value; /* Not used for flag attributes */
size_t value_size; /* Max size of value for output, length for input */
size_t *value_length; /* Returned length of value for output */
} OPENSSL_ITEM;
/* When changing the CRYPTO_LOCK_* list, be sure to maintin the text lock
* names in cryptlib.c
*/
#define CRYPTO_LOCK_ERR 1
#define CRYPTO_LOCK_EX_DATA 2
#define CRYPTO_LOCK_X509 3
#define CRYPTO_LOCK_X509_INFO 4
#define CRYPTO_LOCK_X509_PKEY 5
#define CRYPTO_LOCK_X509_CRL 6
#define CRYPTO_LOCK_X509_REQ 7
#define CRYPTO_LOCK_DSA 8
#define CRYPTO_LOCK_RSA 9
#define CRYPTO_LOCK_EVP_PKEY 10
#define CRYPTO_LOCK_X509_STORE 11
#define CRYPTO_LOCK_SSL_CTX 12
#define CRYPTO_LOCK_SSL_CERT 13
#define CRYPTO_LOCK_SSL_SESSION 14
#define CRYPTO_LOCK_SSL_SESS_CERT 15
#define CRYPTO_LOCK_SSL 16
#define CRYPTO_LOCK_SSL_METHOD 17
#define CRYPTO_LOCK_RAND 18
#define CRYPTO_LOCK_RAND2 19
#define CRYPTO_LOCK_MALLOC 20
#define CRYPTO_LOCK_BIO 21
#define CRYPTO_LOCK_GETHOSTBYNAME 22
#define CRYPTO_LOCK_GETSERVBYNAME 23
#define CRYPTO_LOCK_READDIR 24
#define CRYPTO_LOCK_RSA_BLINDING 25
#define CRYPTO_LOCK_DH 26
#define CRYPTO_LOCK_MALLOC2 27
#define CRYPTO_LOCK_DSO 28
#define CRYPTO_LOCK_DYNLOCK 29
#define CRYPTO_LOCK_ENGINE 30
#define CRYPTO_LOCK_UI 31
#define CRYPTO_LOCK_ECDSA 32
#define CRYPTO_LOCK_EC 33
#define CRYPTO_LOCK_ECDH 34
#define CRYPTO_LOCK_BN 35
#define CRYPTO_LOCK_EC_PRE_COMP 36
#define CRYPTO_LOCK_STORE 37
#define CRYPTO_LOCK_COMP 38
#define CRYPTO_LOCK_FIPS 39
#define CRYPTO_LOCK_FIPS2 40
#define CRYPTO_NUM_LOCKS 41
#define CRYPTO_LOCK 1
#define CRYPTO_UNLOCK 2
#define CRYPTO_READ 4
#define CRYPTO_WRITE 8
#ifndef OPENSSL_NO_LOCKING
#ifndef CRYPTO_w_lock
#define CRYPTO_w_lock(type) \
CRYPTO_lock(CRYPTO_LOCK|CRYPTO_WRITE,type,__FILE__,__LINE__)
#define CRYPTO_w_unlock(type) \
CRYPTO_lock(CRYPTO_UNLOCK|CRYPTO_WRITE,type,__FILE__,__LINE__)
#define CRYPTO_r_lock(type) \
CRYPTO_lock(CRYPTO_LOCK|CRYPTO_READ,type,__FILE__,__LINE__)
#define CRYPTO_r_unlock(type) \
CRYPTO_lock(CRYPTO_UNLOCK|CRYPTO_READ,type,__FILE__,__LINE__)
#define CRYPTO_add(addr,amount,type) \
CRYPTO_add_lock(addr,amount,type,__FILE__,__LINE__)
#endif
#else
#define CRYPTO_w_lock(a)
#define CRYPTO_w_unlock(a)
#define CRYPTO_r_lock(a)
#define CRYPTO_r_unlock(a)
#define CRYPTO_add(a,b,c) ((*(a))+=(b))
#endif
/* Some applications as well as some parts of OpenSSL need to allocate
and deallocate locks in a dynamic fashion. The following typedef
makes this possible in a type-safe manner. */
/* struct CRYPTO_dynlock_value has to be defined by the application. */
typedef struct
{
int references;
struct CRYPTO_dynlock_value *data;
} CRYPTO_dynlock;
/* The following can be used to detect memory leaks in the SSLeay library.
* It used, it turns on malloc checking */
#define CRYPTO_MEM_CHECK_OFF 0x0 /* an enume */
#define CRYPTO_MEM_CHECK_ON 0x1 /* a bit */
#define CRYPTO_MEM_CHECK_ENABLE 0x2 /* a bit */
#define CRYPTO_MEM_CHECK_DISABLE 0x3 /* an enume */
/* The following are bit values to turn on or off options connected to the
* malloc checking functionality */
/* Adds time to the memory checking information */
#define V_CRYPTO_MDEBUG_TIME 0x1 /* a bit */
/* Adds thread number to the memory checking information */
#define V_CRYPTO_MDEBUG_THREAD 0x2 /* a bit */
#define V_CRYPTO_MDEBUG_ALL (V_CRYPTO_MDEBUG_TIME | V_CRYPTO_MDEBUG_THREAD)
/* predec of the BIO type */
typedef struct bio_st BIO_dummy;
struct crypto_ex_data_st
{
STACK_OF(void) *sk;
int dummy; /* gcc is screwing up this data structure :-( */
};
DECLARE_STACK_OF(void)
/* This stuff is basically class callback functions
* The current classes are SSL_CTX, SSL, SSL_SESSION, and a few more */
typedef struct crypto_ex_data_func_st
{
long argl; /* Arbitary long */
void *argp; /* Arbitary void * */
CRYPTO_EX_new *new_func;
CRYPTO_EX_free *free_func;
CRYPTO_EX_dup *dup_func;
} CRYPTO_EX_DATA_FUNCS;
DECLARE_STACK_OF(CRYPTO_EX_DATA_FUNCS)
/* Per class, we have a STACK of CRYPTO_EX_DATA_FUNCS for each CRYPTO_EX_DATA
* entry.
*/
#define CRYPTO_EX_INDEX_BIO 0
#define CRYPTO_EX_INDEX_SSL 1
#define CRYPTO_EX_INDEX_SSL_CTX 2
#define CRYPTO_EX_INDEX_SSL_SESSION 3
#define CRYPTO_EX_INDEX_X509_STORE 4
#define CRYPTO_EX_INDEX_X509_STORE_CTX 5
#define CRYPTO_EX_INDEX_RSA 6
#define CRYPTO_EX_INDEX_DSA 7
#define CRYPTO_EX_INDEX_DH 8
#define CRYPTO_EX_INDEX_ENGINE 9
#define CRYPTO_EX_INDEX_X509 10
#define CRYPTO_EX_INDEX_UI 11
#define CRYPTO_EX_INDEX_ECDSA 12
#define CRYPTO_EX_INDEX_ECDH 13
#define CRYPTO_EX_INDEX_COMP 14
#define CRYPTO_EX_INDEX_STORE 15
/* Dynamically assigned indexes start from this value (don't use directly, use
* via CRYPTO_ex_data_new_class). */
#define CRYPTO_EX_INDEX_USER 100
/* This is the default callbacks, but we can have others as well:
* this is needed in Win32 where the application malloc and the
* library malloc may not be the same.
*/
#define CRYPTO_malloc_init() CRYPTO_set_mem_functions(\
malloc, realloc, free)
#if defined CRYPTO_MDEBUG_ALL || defined CRYPTO_MDEBUG_TIME || defined CRYPTO_MDEBUG_THREAD
# ifndef CRYPTO_MDEBUG /* avoid duplicate #define */
# define CRYPTO_MDEBUG
# endif
#endif
/* Set standard debugging functions (not done by default
* unless CRYPTO_MDEBUG is defined) */
#define CRYPTO_malloc_debug_init() do {\
CRYPTO_set_mem_debug_functions(\
CRYPTO_dbg_malloc,\
CRYPTO_dbg_realloc,\
CRYPTO_dbg_free,\
CRYPTO_dbg_set_options,\
CRYPTO_dbg_get_options);\
} while(0)
int CRYPTO_mem_ctrl(int mode);
int CRYPTO_is_mem_check_on(void);
/* for applications */
#define MemCheck_start() CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ON)
#define MemCheck_stop() CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_OFF)
/* for library-internal use */
#define MemCheck_on() CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_ENABLE)
#define MemCheck_off() CRYPTO_mem_ctrl(CRYPTO_MEM_CHECK_DISABLE)
#define is_MemCheck_on() CRYPTO_is_mem_check_on()
#define OPENSSL_malloc(num) CRYPTO_malloc((int)num,__FILE__,__LINE__)
#define OPENSSL_strdup(str) CRYPTO_strdup((str),__FILE__,__LINE__)
#define OPENSSL_realloc(addr,num) \
CRYPTO_realloc((char *)addr,(int)num,__FILE__,__LINE__)
#define OPENSSL_realloc_clean(addr,old_num,num) \
CRYPTO_realloc_clean(addr,old_num,num,__FILE__,__LINE__)
#define OPENSSL_remalloc(addr,num) \
CRYPTO_remalloc((char **)addr,(int)num,__FILE__,__LINE__)
#define OPENSSL_freeFunc CRYPTO_free
#define OPENSSL_free(addr) CRYPTO_free(addr)
#define OPENSSL_malloc_locked(num) \
CRYPTO_malloc_locked((int)num,__FILE__,__LINE__)
#define OPENSSL_free_locked(addr) CRYPTO_free_locked(addr)
const char *SSLeay_version(int type);
unsigned long SSLeay(void);
int OPENSSL_issetugid(void);
/* An opaque type representing an implementation of "ex_data" support */
typedef struct st_CRYPTO_EX_DATA_IMPL CRYPTO_EX_DATA_IMPL;
/* Return an opaque pointer to the current "ex_data" implementation */
const CRYPTO_EX_DATA_IMPL *CRYPTO_get_ex_data_implementation(void);
/* Sets the "ex_data" implementation to be used (if it's not too late) */
int CRYPTO_set_ex_data_implementation(const CRYPTO_EX_DATA_IMPL *i);
/* Get a new "ex_data" class, and return the corresponding "class_index" */
int CRYPTO_ex_data_new_class(void);
/* Within a given class, get/register a new index */
int CRYPTO_get_ex_new_index(int class_index, long argl, void *argp,
CRYPTO_EX_new *new_func, CRYPTO_EX_dup *dup_func,
CRYPTO_EX_free *free_func);
/* Initialise/duplicate/free CRYPTO_EX_DATA variables corresponding to a given
* class (invokes whatever per-class callbacks are applicable) */
int CRYPTO_new_ex_data(int class_index, void *obj, CRYPTO_EX_DATA *ad);
int CRYPTO_dup_ex_data(int class_index, CRYPTO_EX_DATA *to,
CRYPTO_EX_DATA *from);
void CRYPTO_free_ex_data(int class_index, void *obj, CRYPTO_EX_DATA *ad);
/* Get/set data in a CRYPTO_EX_DATA variable corresponding to a particular index
* (relative to the class type involved) */
int CRYPTO_set_ex_data(CRYPTO_EX_DATA *ad, int idx, void *val);
void *CRYPTO_get_ex_data(const CRYPTO_EX_DATA *ad,int idx);
/* This function cleans up all "ex_data" state. It mustn't be called under
* potential race-conditions. */
void CRYPTO_cleanup_all_ex_data(void);
int CRYPTO_get_new_lockid(char *name);
int CRYPTO_num_locks(void); /* return CRYPTO_NUM_LOCKS (shared libs!) */
void CRYPTO_lock(int mode, int type,const char *file,int line);
void CRYPTO_set_locking_callback(void (*func)(int mode,int type,
const char *file,int line));
void (*CRYPTO_get_locking_callback(void))(int mode,int type,const char *file,
int line);
void CRYPTO_set_add_lock_callback(int (*func)(int *num,int mount,int type,
const char *file, int line));
int (*CRYPTO_get_add_lock_callback(void))(int *num,int mount,int type,
const char *file,int line);
/* Don't use this structure directly. */
typedef struct crypto_threadid_st
{
void *ptr;
unsigned long val;
} CRYPTO_THREADID;
/* Only use CRYPTO_THREADID_set_[numeric|pointer]() within callbacks */
void CRYPTO_THREADID_set_numeric(CRYPTO_THREADID *id, unsigned long val);
void CRYPTO_THREADID_set_pointer(CRYPTO_THREADID *id, void *ptr);
int CRYPTO_THREADID_set_callback(void (*threadid_func)(CRYPTO_THREADID *));
void (*CRYPTO_THREADID_get_callback(void))(CRYPTO_THREADID *);
void CRYPTO_THREADID_current(CRYPTO_THREADID *id);
int CRYPTO_THREADID_cmp(const CRYPTO_THREADID *a, const CRYPTO_THREADID *b);
void CRYPTO_THREADID_cpy(CRYPTO_THREADID *dest, const CRYPTO_THREADID *src);
unsigned long CRYPTO_THREADID_hash(const CRYPTO_THREADID *id);
#ifndef OPENSSL_NO_DEPRECATED
void CRYPTO_set_id_callback(unsigned long (*func)(void));
unsigned long (*CRYPTO_get_id_callback(void))(void);
unsigned long CRYPTO_thread_id(void);
#endif
const char *CRYPTO_get_lock_name(int type);
int CRYPTO_add_lock(int *pointer,int amount,int type, const char *file,
int line);
int CRYPTO_get_new_dynlockid(void);
void CRYPTO_destroy_dynlockid(int i);
struct CRYPTO_dynlock_value *CRYPTO_get_dynlock_value(int i);
void CRYPTO_set_dynlock_create_callback(struct CRYPTO_dynlock_value *(*dyn_create_function)(const char *file, int line));
void CRYPTO_set_dynlock_lock_callback(void (*dyn_lock_function)(int mode, struct CRYPTO_dynlock_value *l, const char *file, int line));
void CRYPTO_set_dynlock_destroy_callback(void (*dyn_destroy_function)(struct CRYPTO_dynlock_value *l, const char *file, int line));
struct CRYPTO_dynlock_value *(*CRYPTO_get_dynlock_create_callback(void))(const char *file,int line);
void (*CRYPTO_get_dynlock_lock_callback(void))(int mode, struct CRYPTO_dynlock_value *l, const char *file,int line);
void (*CRYPTO_get_dynlock_destroy_callback(void))(struct CRYPTO_dynlock_value *l, const char *file,int line);
/* CRYPTO_set_mem_functions includes CRYPTO_set_locked_mem_functions --
* call the latter last if you need different functions */
int CRYPTO_set_mem_functions(void *(*m)(size_t),void *(*r)(void *,size_t), void (*f)(void *));
int CRYPTO_set_locked_mem_functions(void *(*m)(size_t), void (*free_func)(void *));
int CRYPTO_set_mem_ex_functions(void *(*m)(size_t,const char *,int),
void *(*r)(void *,size_t,const char *,int),
void (*f)(void *));
int CRYPTO_set_locked_mem_ex_functions(void *(*m)(size_t,const char *,int),
void (*free_func)(void *));
int CRYPTO_set_mem_debug_functions(void (*m)(void *,int,const char *,int,int),
void (*r)(void *,void *,int,const char *,int,int),
void (*f)(void *,int),
void (*so)(long),
long (*go)(void));
void CRYPTO_get_mem_functions(void *(**m)(size_t),void *(**r)(void *, size_t), void (**f)(void *));
void CRYPTO_get_locked_mem_functions(void *(**m)(size_t), void (**f)(void *));
void CRYPTO_get_mem_ex_functions(void *(**m)(size_t,const char *,int),
void *(**r)(void *, size_t,const char *,int),
void (**f)(void *));
void CRYPTO_get_locked_mem_ex_functions(void *(**m)(size_t,const char *,int),
void (**f)(void *));
void CRYPTO_get_mem_debug_functions(void (**m)(void *,int,const char *,int,int),
void (**r)(void *,void *,int,const char *,int,int),
void (**f)(void *,int),
void (**so)(long),
long (**go)(void));
void *CRYPTO_malloc_locked(int num, const char *file, int line);
void CRYPTO_free_locked(void *ptr);
void *CRYPTO_malloc(int num, const char *file, int line);
char *CRYPTO_strdup(const char *str, const char *file, int line);
void CRYPTO_free(void *ptr);
void *CRYPTO_realloc(void *addr,int num, const char *file, int line);
void *CRYPTO_realloc_clean(void *addr,int old_num,int num,const char *file,
int line);
void *CRYPTO_remalloc(void *addr,int num, const char *file, int line);
void OPENSSL_cleanse(void *ptr, size_t len);
void CRYPTO_set_mem_debug_options(long bits);
long CRYPTO_get_mem_debug_options(void);
#define CRYPTO_push_info(info) \
CRYPTO_push_info_(info, __FILE__, __LINE__);
int CRYPTO_push_info_(const char *info, const char *file, int line);
int CRYPTO_pop_info(void);
int CRYPTO_remove_all_info(void);
/* Default debugging functions (enabled by CRYPTO_malloc_debug_init() macro;
* used as default in CRYPTO_MDEBUG compilations): */
/* The last argument has the following significance:
*
* 0: called before the actual memory allocation has taken place
* 1: called after the actual memory allocation has taken place
*/
void CRYPTO_dbg_malloc(void *addr,int num,const char *file,int line,int before_p);
void CRYPTO_dbg_realloc(void *addr1,void *addr2,int num,const char *file,int line,int before_p);
void CRYPTO_dbg_free(void *addr,int before_p);
/* Tell the debugging code about options. By default, the following values
* apply:
*
* 0: Clear all options.
* V_CRYPTO_MDEBUG_TIME (1): Set the "Show Time" option.
* V_CRYPTO_MDEBUG_THREAD (2): Set the "Show Thread Number" option.
* V_CRYPTO_MDEBUG_ALL (3): 1 + 2
*/
void CRYPTO_dbg_set_options(long bits);
long CRYPTO_dbg_get_options(void);
#ifndef OPENSSL_NO_FP_API
void CRYPTO_mem_leaks_fp(FILE *);
#endif
void CRYPTO_mem_leaks(struct bio_st *bio);
/* unsigned long order, char *file, int line, int num_bytes, char *addr */
typedef void *CRYPTO_MEM_LEAK_CB(unsigned long, const char *, int, int, void *);
void CRYPTO_mem_leaks_cb(CRYPTO_MEM_LEAK_CB *cb);
/* die if we have to */
void OpenSSLDie(const char *file,int line,const char *assertion);
#define OPENSSL_assert(e) (void)((e) ? 0 : (OpenSSLDie(__FILE__, __LINE__, #e),1))
unsigned long *OPENSSL_ia32cap_loc(void);
#define OPENSSL_ia32cap (*(OPENSSL_ia32cap_loc()))
int OPENSSL_isservice(void);
int FIPS_mode(void);
int FIPS_mode_set(int r);
void OPENSSL_init(void);
#define fips_md_init(alg) fips_md_init_ctx(alg, alg)
#ifdef OPENSSL_FIPS
#define fips_md_init_ctx(alg, cx) \
int alg##_Init(cx##_CTX *c) \
{ \
if (FIPS_mode()) OpenSSLDie(__FILE__, __LINE__, \
"Low level API call to digest " #alg " forbidden in FIPS mode!"); \
return private_##alg##_Init(c); \
} \
int private_##alg##_Init(cx##_CTX *c)
#define fips_cipher_abort(alg) \
if (FIPS_mode()) OpenSSLDie(__FILE__, __LINE__, \
"Low level API call to cipher " #alg " forbidden in FIPS mode!")
#else
#define fips_md_init_ctx(alg, cx) \
int alg##_Init(cx##_CTX *c)
#define fips_cipher_abort(alg) while(0)
#endif
/* CRYPTO_memcmp returns zero iff the |len| bytes at |a| and |b| are equal. It
* takes an amount of time dependent on |len|, but independent of the contents
* of |a| and |b|. Unlike memcmp, it cannot be used to put elements into a
* defined order as the return value when a != b is undefined, other than to be
* non-zero. */
int CRYPTO_memcmp(const void *a, const void *b, size_t len);
/* BEGIN ERROR CODES */
/* The following lines are auto generated by the script mkerr.pl. Any changes
* made after this point may be overwritten when the script is next run.
*/
void ERR_load_CRYPTO_strings(void);
/* Error codes for the CRYPTO functions. */
/* Function codes. */
#define CRYPTO_F_CRYPTO_GET_EX_NEW_INDEX 100
#define CRYPTO_F_CRYPTO_GET_NEW_DYNLOCKID 103
#define CRYPTO_F_CRYPTO_GET_NEW_LOCKID 101
#define CRYPTO_F_CRYPTO_SET_EX_DATA 102
#define CRYPTO_F_DEF_ADD_INDEX 104
#define CRYPTO_F_DEF_GET_CLASS 105
#define CRYPTO_F_FIPS_MODE_SET 109
#define CRYPTO_F_INT_DUP_EX_DATA 106
#define CRYPTO_F_INT_FREE_EX_DATA 107
#define CRYPTO_F_INT_NEW_EX_DATA 108
/* Reason codes. */
#define CRYPTO_R_FIPS_MODE_NOT_SUPPORTED 101
#define CRYPTO_R_NO_DYNLOCK_CREATE_CALLBACK 100
#ifdef __cplusplus
}
#endif
#endif
@@ -1,89 +0,0 @@
#ifndef HEADER_OPENSSLV_H
#define HEADER_OPENSSLV_H
/* Numeric release version identifier:
* MNNFFPPS: major minor fix patch status
* The status nibble has one of the values 0 for development, 1 to e for betas
* 1 to 14, and f for release. The patch level is exactly that.
* For example:
* 0.9.3-dev 0x00903000
* 0.9.3-beta1 0x00903001
* 0.9.3-beta2-dev 0x00903002
* 0.9.3-beta2 0x00903002 (same as ...beta2-dev)
* 0.9.3 0x0090300f
* 0.9.3a 0x0090301f
* 0.9.4 0x0090400f
* 1.2.3z 0x102031af
*
* For continuity reasons (because 0.9.5 is already out, and is coded
* 0x00905100), between 0.9.5 and 0.9.6 the coding of the patch level
* part is slightly different, by setting the highest bit. This means
* that 0.9.5a looks like this: 0x0090581f. At 0.9.6, we can start
* with 0x0090600S...
*
* (Prior to 0.9.3-dev a different scheme was used: 0.9.2b is 0x0922.)
* (Prior to 0.9.5a beta1, a different scheme was used: MMNNFFRBB for
* major minor fix final patch/beta)
*/
#define OPENSSL_VERSION_NUMBER 0x1000109fL
#ifdef OPENSSL_FIPS
#define OPENSSL_VERSION_TEXT "OpenSSL 1.0.1i-fips 6 Aug 2014"
#else
#define OPENSSL_VERSION_TEXT "OpenSSL 1.0.1i 6 Aug 2014"
#endif
#define OPENSSL_VERSION_PTEXT " part of " OPENSSL_VERSION_TEXT
/* The macros below are to be used for shared library (.so, .dll, ...)
* versioning. That kind of versioning works a bit differently between
* operating systems. The most usual scheme is to set a major and a minor
* number, and have the runtime loader check that the major number is equal
* to what it was at application link time, while the minor number has to
* be greater or equal to what it was at application link time. With this
* scheme, the version number is usually part of the file name, like this:
*
* libcrypto.so.0.9
*
* Some unixen also make a softlink with the major verson number only:
*
* libcrypto.so.0
*
* On Tru64 and IRIX 6.x it works a little bit differently. There, the
* shared library version is stored in the file, and is actually a series
* of versions, separated by colons. The rightmost version present in the
* library when linking an application is stored in the application to be
* matched at run time. When the application is run, a check is done to
* see if the library version stored in the application matches any of the
* versions in the version string of the library itself.
* This version string can be constructed in any way, depending on what
* kind of matching is desired. However, to implement the same scheme as
* the one used in the other unixen, all compatible versions, from lowest
* to highest, should be part of the string. Consecutive builds would
* give the following versions strings:
*
* 3.0
* 3.0:3.1
* 3.0:3.1:3.2
* 4.0
* 4.0:4.1
*
* Notice how version 4 is completely incompatible with version, and
* therefore give the breach you can see.
*
* There may be other schemes as well that I haven't yet discovered.
*
* So, here's the way it works here: first of all, the library version
* number doesn't need at all to match the overall OpenSSL version.
* However, it's nice and more understandable if it actually does.
* The current library version is stored in the macro SHLIB_VERSION_NUMBER,
* which is just a piece of text in the format "M.m.e" (Major, minor, edit).
* For the sake of Tru64, IRIX, and any other OS that behaves in similar ways,
* we need to keep a history of version numbers, which is done in the
* macro SHLIB_VERSION_HISTORY. The numbers are separated by colons and
* should only keep the versions that are binary compatible with the current.
*/
#define SHLIB_VERSION_HISTORY ""
#define SHLIB_VERSION_NUMBER "1.0.0"
#endif /* HEADER_OPENSSLV_H */
@@ -1,202 +0,0 @@
/* ====================================================================
* Copyright (c) 1998-2001 The OpenSSL Project. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* 3. All advertising materials mentioning features or use of this
* software must display the following acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
*
* 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
* endorse or promote products derived from this software without
* prior written permission. For written permission, please contact
* openssl-core@openssl.org.
*
* 5. Products derived from this software may not be called "OpenSSL"
* nor may "OpenSSL" appear in their names without prior written
* permission of the OpenSSL Project.
*
* 6. Redistributions of any form whatsoever must retain the following
* acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit (http://www.openssl.org/)"
*
* THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
* EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
* ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
* OF THE POSSIBILITY OF SUCH DAMAGE.
* ====================================================================
*
* This product includes cryptographic software written by Eric Young
* (eay@cryptsoft.com). This product includes software written by Tim
* Hudson (tjh@cryptsoft.com).
*
*/
#ifndef HEADER_OPENSSL_TYPES_H
#define HEADER_OPENSSL_TYPES_H
#include <openssl/e_os2.h>
#ifdef NO_ASN1_TYPEDEFS
#define ASN1_INTEGER ASN1_STRING
#define ASN1_ENUMERATED ASN1_STRING
#define ASN1_BIT_STRING ASN1_STRING
#define ASN1_OCTET_STRING ASN1_STRING
#define ASN1_PRINTABLESTRING ASN1_STRING
#define ASN1_T61STRING ASN1_STRING
#define ASN1_IA5STRING ASN1_STRING
#define ASN1_UTCTIME ASN1_STRING
#define ASN1_GENERALIZEDTIME ASN1_STRING
#define ASN1_TIME ASN1_STRING
#define ASN1_GENERALSTRING ASN1_STRING
#define ASN1_UNIVERSALSTRING ASN1_STRING
#define ASN1_BMPSTRING ASN1_STRING
#define ASN1_VISIBLESTRING ASN1_STRING
#define ASN1_UTF8STRING ASN1_STRING
#define ASN1_BOOLEAN int
#define ASN1_NULL int
#else
typedef struct asn1_string_st ASN1_INTEGER;
typedef struct asn1_string_st ASN1_ENUMERATED;
typedef struct asn1_string_st ASN1_BIT_STRING;
typedef struct asn1_string_st ASN1_OCTET_STRING;
typedef struct asn1_string_st ASN1_PRINTABLESTRING;
typedef struct asn1_string_st ASN1_T61STRING;
typedef struct asn1_string_st ASN1_IA5STRING;
typedef struct asn1_string_st ASN1_GENERALSTRING;
typedef struct asn1_string_st ASN1_UNIVERSALSTRING;
typedef struct asn1_string_st ASN1_BMPSTRING;
typedef struct asn1_string_st ASN1_UTCTIME;
typedef struct asn1_string_st ASN1_TIME;
typedef struct asn1_string_st ASN1_GENERALIZEDTIME;
typedef struct asn1_string_st ASN1_VISIBLESTRING;
typedef struct asn1_string_st ASN1_UTF8STRING;
typedef struct asn1_string_st ASN1_STRING;
typedef int ASN1_BOOLEAN;
typedef int ASN1_NULL;
#endif
typedef struct ASN1_ITEM_st ASN1_ITEM;
typedef struct asn1_pctx_st ASN1_PCTX;
#ifdef OPENSSL_SYS_WIN32
#undef X509_NAME
#undef X509_EXTENSIONS
#undef X509_CERT_PAIR
#undef PKCS7_ISSUER_AND_SERIAL
#undef OCSP_REQUEST
#undef OCSP_RESPONSE
#endif
#ifdef BIGNUM
#undef BIGNUM
#endif
typedef struct bignum_st BIGNUM;
typedef struct bignum_ctx BN_CTX;
typedef struct bn_blinding_st BN_BLINDING;
typedef struct bn_mont_ctx_st BN_MONT_CTX;
typedef struct bn_recp_ctx_st BN_RECP_CTX;
typedef struct bn_gencb_st BN_GENCB;
typedef struct buf_mem_st BUF_MEM;
typedef struct evp_cipher_st EVP_CIPHER;
typedef struct evp_cipher_ctx_st EVP_CIPHER_CTX;
typedef struct env_md_st EVP_MD;
typedef struct env_md_ctx_st EVP_MD_CTX;
typedef struct evp_pkey_st EVP_PKEY;
typedef struct evp_pkey_asn1_method_st EVP_PKEY_ASN1_METHOD;
typedef struct evp_pkey_method_st EVP_PKEY_METHOD;
typedef struct evp_pkey_ctx_st EVP_PKEY_CTX;
typedef struct dh_st DH;
typedef struct dh_method DH_METHOD;
typedef struct dsa_st DSA;
typedef struct dsa_method DSA_METHOD;
typedef struct rsa_st RSA;
typedef struct rsa_meth_st RSA_METHOD;
typedef struct rand_meth_st RAND_METHOD;
typedef struct ecdh_method ECDH_METHOD;
typedef struct ecdsa_method ECDSA_METHOD;
typedef struct x509_st X509;
typedef struct X509_algor_st X509_ALGOR;
typedef struct X509_crl_st X509_CRL;
typedef struct x509_crl_method_st X509_CRL_METHOD;
typedef struct x509_revoked_st X509_REVOKED;
typedef struct X509_name_st X509_NAME;
typedef struct X509_pubkey_st X509_PUBKEY;
typedef struct x509_store_st X509_STORE;
typedef struct x509_store_ctx_st X509_STORE_CTX;
typedef struct pkcs8_priv_key_info_st PKCS8_PRIV_KEY_INFO;
typedef struct v3_ext_ctx X509V3_CTX;
typedef struct conf_st CONF;
typedef struct store_st STORE;
typedef struct store_method_st STORE_METHOD;
typedef struct ui_st UI;
typedef struct ui_method_st UI_METHOD;
typedef struct st_ERR_FNS ERR_FNS;
typedef struct engine_st ENGINE;
typedef struct ssl_st SSL;
typedef struct ssl_ctx_st SSL_CTX;
typedef struct X509_POLICY_NODE_st X509_POLICY_NODE;
typedef struct X509_POLICY_LEVEL_st X509_POLICY_LEVEL;
typedef struct X509_POLICY_TREE_st X509_POLICY_TREE;
typedef struct X509_POLICY_CACHE_st X509_POLICY_CACHE;
typedef struct AUTHORITY_KEYID_st AUTHORITY_KEYID;
typedef struct DIST_POINT_st DIST_POINT;
typedef struct ISSUING_DIST_POINT_st ISSUING_DIST_POINT;
typedef struct NAME_CONSTRAINTS_st NAME_CONSTRAINTS;
/* If placed in pkcs12.h, we end up with a circular depency with pkcs7.h */
#define DECLARE_PKCS12_STACK_OF(type) /* Nothing */
#define IMPLEMENT_PKCS12_STACK_OF(type) /* Nothing */
typedef struct crypto_ex_data_st CRYPTO_EX_DATA;
/* Callback types for crypto.h */
typedef int CRYPTO_EX_new(void *parent, void *ptr, CRYPTO_EX_DATA *ad,
int idx, long argl, void *argp);
typedef void CRYPTO_EX_free(void *parent, void *ptr, CRYPTO_EX_DATA *ad,
int idx, long argl, void *argp);
typedef int CRYPTO_EX_dup(CRYPTO_EX_DATA *to, CRYPTO_EX_DATA *from, void *from_d,
int idx, long argl, void *argp);
typedef struct ocsp_req_ctx_st OCSP_REQ_CTX;
typedef struct ocsp_response_st OCSP_RESPONSE;
typedef struct ocsp_responder_id_st OCSP_RESPID;
#endif /* def HEADER_OPENSSL_TYPES_H */
File diff suppressed because it is too large. Load diff
@@ -1,108 +0,0 @@
/* crypto/stack/stack.h */
/* Copyright (C) 1995-1998 Eric Young (eay@cryptsoft.com)
* All rights reserved.
*
* This package is an SSL implementation written
* by Eric Young (eay@cryptsoft.com).
* The implementation was written so as to conform with Netscapes SSL.
*
* This library is free for commercial and non-commercial use as long as
* the following conditions are aheared to. The following conditions
* apply to all code found in this distribution, be it the RC4, RSA,
* lhash, DES, etc., code; not just the SSL code. The SSL documentation
* included with this distribution is covered by the same copyright terms
* except that the holder is Tim Hudson (tjh@cryptsoft.com).
*
* Copyright remains Eric Young's, and as such any Copyright notices in
* the code are not to be removed.
* If this package is used in a product, Eric Young should be given attribution
* as the author of the parts of the library used.
* This can be in the form of a textual message at program startup or
* in documentation (online or textual) provided with the package.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* "This product includes cryptographic software written by
* Eric Young (eay@cryptsoft.com)"
* The word 'cryptographic' can be left out if the rouines from the library
* being used are not cryptographic related :-).
* 4. If you include any Windows specific code (or a derivative thereof) from
* the apps directory (application code) you must include an acknowledgement:
* "This product includes software written by Tim Hudson (tjh@cryptsoft.com)"
*
* THIS SOFTWARE IS PROVIDED BY ERIC YOUNG ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* The licence and distribution terms for any publically available version or
* derivative of this code cannot be changed. i.e. this code cannot simply be
* copied and put under another distribution licence
* [including the GNU Public Licence.]
*/
#ifndef HEADER_STACK_H
#define HEADER_STACK_H
#ifdef __cplusplus
extern "C" {
#endif
typedef struct stack_st
{
int num;
char **data;
int sorted;
int num_alloc;
int (*comp)(const void *, const void *);
} _STACK; /* Use STACK_OF(...) instead */
#define M_sk_num(sk) ((sk) ? (sk)->num:-1)
#define M_sk_value(sk,n) ((sk) ? (sk)->data[n] : NULL)
int sk_num(const _STACK *);
void *sk_value(const _STACK *, int);
void *sk_set(_STACK *, int, void *);
_STACK *sk_new(int (*cmp)(const void *, const void *));
_STACK *sk_new_null(void);
void sk_free(_STACK *);
void sk_pop_free(_STACK *st, void (*func)(void *));
int sk_insert(_STACK *sk, void *data, int where);
void *sk_delete(_STACK *st, int loc);
void *sk_delete_ptr(_STACK *st, void *p);
int sk_find(_STACK *st, void *data);
int sk_find_ex(_STACK *st, void *data);
int sk_push(_STACK *st, void *data);
int sk_unshift(_STACK *st, void *data);
void *sk_shift(_STACK *st);
void *sk_pop(_STACK *st);
void sk_zero(_STACK *st);
int (*sk_set_cmp_func(_STACK *sk, int (*c)(const void *, const void *)))
(const void *, const void *);
_STACK *sk_dup(_STACK *st);
void sk_sort(_STACK *st);
int sk_is_sorted(const _STACK *st);
#ifdef __cplusplus
}
#endif
#endif
@@ -1,481 +0,0 @@
/* ====================================================================
* Copyright (c) 1999 The OpenSSL Project. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in
* the documentation and/or other materials provided with the
* distribution.
*
* 3. All advertising materials mentioning features or use of this
* software must display the following acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit. (http://www.openssl.org/)"
*
* 4. The names "OpenSSL Toolkit" and "OpenSSL Project" must not be used to
* endorse or promote products derived from this software without
* prior written permission. For written permission, please contact
* openssl-core@openssl.org.
*
* 5. Products derived from this software may not be called "OpenSSL"
* nor may "OpenSSL" appear in their names without prior written
* permission of the OpenSSL Project.
*
* 6. Redistributions of any form whatsoever must retain the following
* acknowledgment:
* "This product includes software developed by the OpenSSL Project
* for use in the OpenSSL Toolkit (http://www.openssl.org/)"
*
* THIS SOFTWARE IS PROVIDED BY THE OpenSSL PROJECT ``AS IS'' AND ANY
* EXPRESSED OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR
* PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE OpenSSL PROJECT OR
* ITS CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
* SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
* LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
* STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
* OF THE POSSIBILITY OF SUCH DAMAGE.
* ====================================================================
*
* This product includes cryptographic software written by Eric Young
* (eay@cryptsoft.com). This product includes software written by Tim
* Hudson (tjh@cryptsoft.com).
*
*/
#ifndef HEADER_SYMHACKS_H
#define HEADER_SYMHACKS_H
#include <openssl/e_os2.h>
/* Hacks to solve the problem with linkers incapable of handling very long
symbol names. In the case of VMS, the limit is 31 characters on VMS for
VAX. */
/* Note that this affects util/libeay.num and util/ssleay.num... you may
change those manually, but that's not recommended, as those files are
controlled centrally and updated on Unix, and the central definition
may disagree with yours, which in turn may come with shareable library
incompatibilities. */
#ifdef OPENSSL_SYS_VMS
/* Hack a long name in crypto/ex_data.c */
#undef CRYPTO_get_ex_data_implementation
#define CRYPTO_get_ex_data_implementation CRYPTO_get_ex_data_impl
#undef CRYPTO_set_ex_data_implementation
#define CRYPTO_set_ex_data_implementation CRYPTO_set_ex_data_impl
/* Hack a long name in crypto/asn1/a_mbstr.c */
#undef ASN1_STRING_set_default_mask_asc
#define ASN1_STRING_set_default_mask_asc ASN1_STRING_set_def_mask_asc
#if 0 /* No longer needed, since safestack macro magic does the job */
/* Hack the names created with DECLARE_ASN1_SET_OF(PKCS7_SIGNER_INFO) */
#undef i2d_ASN1_SET_OF_PKCS7_SIGNER_INFO
#define i2d_ASN1_SET_OF_PKCS7_SIGNER_INFO i2d_ASN1_SET_OF_PKCS7_SIGINF
#undef d2i_ASN1_SET_OF_PKCS7_SIGNER_INFO
#define d2i_ASN1_SET_OF_PKCS7_SIGNER_INFO d2i_ASN1_SET_OF_PKCS7_SIGINF
#endif
#if 0 /* No longer needed, since safestack macro magic does the job */
/* Hack the names created with DECLARE_ASN1_SET_OF(PKCS7_RECIP_INFO) */
#undef i2d_ASN1_SET_OF_PKCS7_RECIP_INFO
#define i2d_ASN1_SET_OF_PKCS7_RECIP_INFO i2d_ASN1_SET_OF_PKCS7_RECINF
#undef d2i_ASN1_SET_OF_PKCS7_RECIP_INFO
#define d2i_ASN1_SET_OF_PKCS7_RECIP_INFO d2i_ASN1_SET_OF_PKCS7_RECINF
#endif
#if 0 /* No longer needed, since safestack macro magic does the job */
/* Hack the names created with DECLARE_ASN1_SET_OF(ACCESS_DESCRIPTION) */
#undef i2d_ASN1_SET_OF_ACCESS_DESCRIPTION
#define i2d_ASN1_SET_OF_ACCESS_DESCRIPTION i2d_ASN1_SET_OF_ACC_DESC
#undef d2i_ASN1_SET_OF_ACCESS_DESCRIPTION
#define d2i_ASN1_SET_OF_ACCESS_DESCRIPTION d2i_ASN1_SET_OF_ACC_DESC
#endif
/* Hack the names created with DECLARE_PEM_rw(NETSCAPE_CERT_SEQUENCE) */
#undef PEM_read_NETSCAPE_CERT_SEQUENCE
#define PEM_read_NETSCAPE_CERT_SEQUENCE PEM_read_NS_CERT_SEQ
#undef PEM_write_NETSCAPE_CERT_SEQUENCE
#define PEM_write_NETSCAPE_CERT_SEQUENCE PEM_write_NS_CERT_SEQ
#undef PEM_read_bio_NETSCAPE_CERT_SEQUENCE
#define PEM_read_bio_NETSCAPE_CERT_SEQUENCE PEM_read_bio_NS_CERT_SEQ
#undef PEM_write_bio_NETSCAPE_CERT_SEQUENCE
#define PEM_write_bio_NETSCAPE_CERT_SEQUENCE PEM_write_bio_NS_CERT_SEQ
#undef PEM_write_cb_bio_NETSCAPE_CERT_SEQUENCE
#define PEM_write_cb_bio_NETSCAPE_CERT_SEQUENCE PEM_write_cb_bio_NS_CERT_SEQ
/* Hack the names created with DECLARE_PEM_rw(PKCS8_PRIV_KEY_INFO) */
#undef PEM_read_PKCS8_PRIV_KEY_INFO
#define PEM_read_PKCS8_PRIV_KEY_INFO PEM_read_P8_PRIV_KEY_INFO
#undef PEM_write_PKCS8_PRIV_KEY_INFO
#define PEM_write_PKCS8_PRIV_KEY_INFO PEM_write_P8_PRIV_KEY_INFO
#undef PEM_read_bio_PKCS8_PRIV_KEY_INFO
#define PEM_read_bio_PKCS8_PRIV_KEY_INFO PEM_read_bio_P8_PRIV_KEY_INFO
#undef PEM_write_bio_PKCS8_PRIV_KEY_INFO
#define PEM_write_bio_PKCS8_PRIV_KEY_INFO PEM_write_bio_P8_PRIV_KEY_INFO
#undef PEM_write_cb_bio_PKCS8_PRIV_KEY_INFO
#define PEM_write_cb_bio_PKCS8_PRIV_KEY_INFO PEM_wrt_cb_bio_P8_PRIV_KEY_INFO
/* Hack other PEM names */
#undef PEM_write_bio_PKCS8PrivateKey_nid
#define PEM_write_bio_PKCS8PrivateKey_nid PEM_write_bio_PKCS8PrivKey_nid
/* Hack some long X509 names */
#undef X509_REVOKED_get_ext_by_critical
#define X509_REVOKED_get_ext_by_critical X509_REVOKED_get_ext_by_critic
#undef X509_policy_tree_get0_user_policies
#define X509_policy_tree_get0_user_policies X509_pcy_tree_get0_usr_policies
#undef X509_policy_node_get0_qualifiers
#define X509_policy_node_get0_qualifiers X509_pcy_node_get0_qualifiers
#undef X509_STORE_CTX_get_explicit_policy
#define X509_STORE_CTX_get_explicit_policy X509_STORE_CTX_get_expl_policy
#undef X509_STORE_CTX_get0_current_issuer
#define X509_STORE_CTX_get0_current_issuer X509_STORE_CTX_get0_cur_issuer
/* Hack some long CRYPTO names */
#undef CRYPTO_set_dynlock_destroy_callback
#define CRYPTO_set_dynlock_destroy_callback CRYPTO_set_dynlock_destroy_cb
#undef CRYPTO_set_dynlock_create_callback
#define CRYPTO_set_dynlock_create_callback CRYPTO_set_dynlock_create_cb
#undef CRYPTO_set_dynlock_lock_callback
#define CRYPTO_set_dynlock_lock_callback CRYPTO_set_dynlock_lock_cb
#undef CRYPTO_get_dynlock_lock_callback
#define CRYPTO_get_dynlock_lock_callback CRYPTO_get_dynlock_lock_cb
#undef CRYPTO_get_dynlock_destroy_callback
#define CRYPTO_get_dynlock_destroy_callback CRYPTO_get_dynlock_destroy_cb
#undef CRYPTO_get_dynlock_create_callback
#define CRYPTO_get_dynlock_create_callback CRYPTO_get_dynlock_create_cb
#undef CRYPTO_set_locked_mem_ex_functions
#define CRYPTO_set_locked_mem_ex_functions CRYPTO_set_locked_mem_ex_funcs
#undef CRYPTO_get_locked_mem_ex_functions
#define CRYPTO_get_locked_mem_ex_functions CRYPTO_get_locked_mem_ex_funcs
/* Hack some long SSL names */
#undef SSL_CTX_set_default_verify_paths
#define SSL_CTX_set_default_verify_paths SSL_CTX_set_def_verify_paths
#undef SSL_get_ex_data_X509_STORE_CTX_idx
#define SSL_get_ex_data_X509_STORE_CTX_idx SSL_get_ex_d_X509_STORE_CTX_idx
#undef SSL_add_file_cert_subjects_to_stack
#define SSL_add_file_cert_subjects_to_stack SSL_add_file_cert_subjs_to_stk
#undef SSL_add_dir_cert_subjects_to_stack
#define SSL_add_dir_cert_subjects_to_stack SSL_add_dir_cert_subjs_to_stk
#undef SSL_CTX_use_certificate_chain_file
#define SSL_CTX_use_certificate_chain_file SSL_CTX_use_cert_chain_file
#undef SSL_CTX_set_cert_verify_callback
#define SSL_CTX_set_cert_verify_callback SSL_CTX_set_cert_verify_cb
#undef SSL_CTX_set_default_passwd_cb_userdata
#define SSL_CTX_set_default_passwd_cb_userdata SSL_CTX_set_def_passwd_cb_ud
#undef SSL_COMP_get_compression_methods
#define SSL_COMP_get_compression_methods SSL_COMP_get_compress_methods
#undef ssl_add_clienthello_renegotiate_ext
#define ssl_add_clienthello_renegotiate_ext ssl_add_clienthello_reneg_ext
#undef ssl_add_serverhello_renegotiate_ext
#define ssl_add_serverhello_renegotiate_ext ssl_add_serverhello_reneg_ext
#undef ssl_parse_clienthello_renegotiate_ext
#define ssl_parse_clienthello_renegotiate_ext ssl_parse_clienthello_reneg_ext
#undef ssl_parse_serverhello_renegotiate_ext
#define ssl_parse_serverhello_renegotiate_ext ssl_parse_serverhello_reneg_ext
#undef SSL_srp_server_param_with_username
#define SSL_srp_server_param_with_username SSL_srp_server_param_with_un
#undef SSL_CTX_set_srp_client_pwd_callback
#define SSL_CTX_set_srp_client_pwd_callback SSL_CTX_set_srp_client_pwd_cb
#undef SSL_CTX_set_srp_verify_param_callback
#define SSL_CTX_set_srp_verify_param_callback SSL_CTX_set_srp_vfy_param_cb
#undef SSL_CTX_set_srp_username_callback
#define SSL_CTX_set_srp_username_callback SSL_CTX_set_srp_un_cb
#undef ssl_add_clienthello_use_srtp_ext
#define ssl_add_clienthello_use_srtp_ext ssl_add_clihello_use_srtp_ext
#undef ssl_add_serverhello_use_srtp_ext
#define ssl_add_serverhello_use_srtp_ext ssl_add_serhello_use_srtp_ext
#undef ssl_parse_clienthello_use_srtp_ext
#define ssl_parse_clienthello_use_srtp_ext ssl_parse_clihello_use_srtp_ext
#undef ssl_parse_serverhello_use_srtp_ext
#define ssl_parse_serverhello_use_srtp_ext ssl_parse_serhello_use_srtp_ext
#undef SSL_CTX_set_next_protos_advertised_cb
#define SSL_CTX_set_next_protos_advertised_cb SSL_CTX_set_next_protos_adv_cb
#undef SSL_CTX_set_next_proto_select_cb
#define SSL_CTX_set_next_proto_select_cb SSL_CTX_set_next_proto_sel_cb
#undef ssl3_cbc_record_digest_supported
#define ssl3_cbc_record_digest_supported ssl3_cbc_record_digest_support
#undef ssl_check_clienthello_tlsext_late
#define ssl_check_clienthello_tlsext_late ssl_check_clihello_tlsext_late
#undef ssl_check_clienthello_tlsext_early
#define ssl_check_clienthello_tlsext_early ssl_check_clihello_tlsext_early
/* Hack some long ENGINE names */
#undef ENGINE_get_default_BN_mod_exp_crt
#define ENGINE_get_default_BN_mod_exp_crt ENGINE_get_def_BN_mod_exp_crt
#undef ENGINE_set_default_BN_mod_exp_crt
#define ENGINE_set_default_BN_mod_exp_crt ENGINE_set_def_BN_mod_exp_crt
#undef ENGINE_set_load_privkey_function
#define ENGINE_set_load_privkey_function ENGINE_set_load_privkey_fn
#undef ENGINE_get_load_privkey_function
#define ENGINE_get_load_privkey_function ENGINE_get_load_privkey_fn
#undef ENGINE_unregister_pkey_asn1_meths
#define ENGINE_unregister_pkey_asn1_meths ENGINE_unreg_pkey_asn1_meths
#undef ENGINE_register_all_pkey_asn1_meths
#define ENGINE_register_all_pkey_asn1_meths ENGINE_reg_all_pkey_asn1_meths
#undef ENGINE_set_default_pkey_asn1_meths
#define ENGINE_set_default_pkey_asn1_meths ENGINE_set_def_pkey_asn1_meths
#undef ENGINE_get_pkey_asn1_meth_engine
#define ENGINE_get_pkey_asn1_meth_engine ENGINE_get_pkey_asn1_meth_eng
#undef ENGINE_set_load_ssl_client_cert_function
#define ENGINE_set_load_ssl_client_cert_function \
ENGINE_set_ld_ssl_clnt_cert_fn
#undef ENGINE_get_ssl_client_cert_function
#define ENGINE_get_ssl_client_cert_function ENGINE_get_ssl_client_cert_fn
/* Hack some long OCSP names */
#undef OCSP_REQUEST_get_ext_by_critical
#define OCSP_REQUEST_get_ext_by_critical OCSP_REQUEST_get_ext_by_crit
#undef OCSP_BASICRESP_get_ext_by_critical
#define OCSP_BASICRESP_get_ext_by_critical OCSP_BASICRESP_get_ext_by_crit
#undef OCSP_SINGLERESP_get_ext_by_critical
#define OCSP_SINGLERESP_get_ext_by_critical OCSP_SINGLERESP_get_ext_by_crit
/* Hack some long DES names */
#undef _ossl_old_des_ede3_cfb64_encrypt
#define _ossl_old_des_ede3_cfb64_encrypt _ossl_odes_ede3_cfb64_encrypt
#undef _ossl_old_des_ede3_ofb64_encrypt
#define _ossl_old_des_ede3_ofb64_encrypt _ossl_odes_ede3_ofb64_encrypt
/* Hack some long EVP names */
#undef OPENSSL_add_all_algorithms_noconf
#define OPENSSL_add_all_algorithms_noconf OPENSSL_add_all_algo_noconf
#undef OPENSSL_add_all_algorithms_conf
#define OPENSSL_add_all_algorithms_conf OPENSSL_add_all_algo_conf
#undef EVP_PKEY_meth_set_verify_recover
#define EVP_PKEY_meth_set_verify_recover EVP_PKEY_meth_set_vrfy_recover
/* Hack some long EC names */
#undef EC_GROUP_set_point_conversion_form
#define EC_GROUP_set_point_conversion_form EC_GROUP_set_point_conv_form
#undef EC_GROUP_get_point_conversion_form
#define EC_GROUP_get_point_conversion_form EC_GROUP_get_point_conv_form
#undef EC_GROUP_clear_free_all_extra_data
#define EC_GROUP_clear_free_all_extra_data EC_GROUP_clr_free_all_xtra_data
#undef EC_KEY_set_public_key_affine_coordinates
#define EC_KEY_set_public_key_affine_coordinates \
EC_KEY_set_pub_key_aff_coords
#undef EC_POINT_set_Jprojective_coordinates_GFp
#define EC_POINT_set_Jprojective_coordinates_GFp \
EC_POINT_set_Jproj_coords_GFp
#undef EC_POINT_get_Jprojective_coordinates_GFp
#define EC_POINT_get_Jprojective_coordinates_GFp \
EC_POINT_get_Jproj_coords_GFp
#undef EC_POINT_set_affine_coordinates_GFp
#define EC_POINT_set_affine_coordinates_GFp EC_POINT_set_affine_coords_GFp
#undef EC_POINT_get_affine_coordinates_GFp
#define EC_POINT_get_affine_coordinates_GFp EC_POINT_get_affine_coords_GFp
#undef EC_POINT_set_compressed_coordinates_GFp
#define EC_POINT_set_compressed_coordinates_GFp EC_POINT_set_compr_coords_GFp
#undef EC_POINT_set_affine_coordinates_GF2m
#define EC_POINT_set_affine_coordinates_GF2m EC_POINT_set_affine_coords_GF2m
#undef EC_POINT_get_affine_coordinates_GF2m
#define EC_POINT_get_affine_coordinates_GF2m EC_POINT_get_affine_coords_GF2m
#undef EC_POINT_set_compressed_coordinates_GF2m
#define EC_POINT_set_compressed_coordinates_GF2m \
EC_POINT_set_compr_coords_GF2m
#undef ec_GF2m_simple_group_clear_finish
#define ec_GF2m_simple_group_clear_finish ec_GF2m_simple_grp_clr_finish
#undef ec_GF2m_simple_group_check_discriminant
#define ec_GF2m_simple_group_check_discriminant ec_GF2m_simple_grp_chk_discrim
#undef ec_GF2m_simple_point_clear_finish
#define ec_GF2m_simple_point_clear_finish ec_GF2m_simple_pt_clr_finish
#undef ec_GF2m_simple_point_set_to_infinity
#define ec_GF2m_simple_point_set_to_infinity ec_GF2m_simple_pt_set_to_inf
#undef ec_GF2m_simple_points_make_affine
#define ec_GF2m_simple_points_make_affine ec_GF2m_simple_pts_make_affine
#undef ec_GF2m_simple_point_set_affine_coordinates
#define ec_GF2m_simple_point_set_affine_coordinates \
ec_GF2m_smp_pt_set_af_coords
#undef ec_GF2m_simple_point_get_affine_coordinates
#define ec_GF2m_simple_point_get_affine_coordinates \
ec_GF2m_smp_pt_get_af_coords
#undef ec_GF2m_simple_set_compressed_coordinates
#define ec_GF2m_simple_set_compressed_coordinates \
ec_GF2m_smp_set_compr_coords
#undef ec_GFp_simple_group_set_curve_GFp
#define ec_GFp_simple_group_set_curve_GFp ec_GFp_simple_grp_set_curve_GFp
#undef ec_GFp_simple_group_get_curve_GFp
#define ec_GFp_simple_group_get_curve_GFp ec_GFp_simple_grp_get_curve_GFp
#undef ec_GFp_simple_group_clear_finish
#define ec_GFp_simple_group_clear_finish ec_GFp_simple_grp_clear_finish
#undef ec_GFp_simple_group_set_generator
#define ec_GFp_simple_group_set_generator ec_GFp_simple_grp_set_generator
#undef ec_GFp_simple_group_get0_generator
#define ec_GFp_simple_group_get0_generator ec_GFp_simple_grp_gt0_generator
#undef ec_GFp_simple_group_get_cofactor
#define ec_GFp_simple_group_get_cofactor ec_GFp_simple_grp_get_cofactor
#undef ec_GFp_simple_point_clear_finish
#define ec_GFp_simple_point_clear_finish ec_GFp_simple_pt_clear_finish
#undef ec_GFp_simple_point_set_to_infinity
#define ec_GFp_simple_point_set_to_infinity ec_GFp_simple_pt_set_to_inf
#undef ec_GFp_simple_points_make_affine
#define ec_GFp_simple_points_make_affine ec_GFp_simple_pts_make_affine
#undef ec_GFp_simple_set_Jprojective_coordinates_GFp
#define ec_GFp_simple_set_Jprojective_coordinates_GFp \
ec_GFp_smp_set_Jproj_coords_GFp
#undef ec_GFp_simple_get_Jprojective_coordinates_GFp
#define ec_GFp_simple_get_Jprojective_coordinates_GFp \
ec_GFp_smp_get_Jproj_coords_GFp
#undef ec_GFp_simple_point_set_affine_coordinates_GFp
#define ec_GFp_simple_point_set_affine_coordinates_GFp \
ec_GFp_smp_pt_set_af_coords_GFp
#undef ec_GFp_simple_point_get_affine_coordinates_GFp
#define ec_GFp_simple_point_get_affine_coordinates_GFp \
ec_GFp_smp_pt_get_af_coords_GFp
#undef ec_GFp_simple_set_compressed_coordinates_GFp
#define ec_GFp_simple_set_compressed_coordinates_GFp \
ec_GFp_smp_set_compr_coords_GFp
#undef ec_GFp_simple_point_set_affine_coordinates
#define ec_GFp_simple_point_set_affine_coordinates \
ec_GFp_smp_pt_set_af_coords
#undef ec_GFp_simple_point_get_affine_coordinates
#define ec_GFp_simple_point_get_affine_coordinates \
ec_GFp_smp_pt_get_af_coords
#undef ec_GFp_simple_set_compressed_coordinates
#define ec_GFp_simple_set_compressed_coordinates \
ec_GFp_smp_set_compr_coords
#undef ec_GFp_simple_group_check_discriminant
#define ec_GFp_simple_group_check_discriminant ec_GFp_simple_grp_chk_discrim
/* Hack som long STORE names */
#undef STORE_method_set_initialise_function
#define STORE_method_set_initialise_function STORE_meth_set_initialise_fn
#undef STORE_method_set_cleanup_function
#define STORE_method_set_cleanup_function STORE_meth_set_cleanup_fn
#undef STORE_method_set_generate_function
#define STORE_method_set_generate_function STORE_meth_set_generate_fn
#undef STORE_method_set_modify_function
#define STORE_method_set_modify_function STORE_meth_set_modify_fn
#undef STORE_method_set_revoke_function
#define STORE_method_set_revoke_function STORE_meth_set_revoke_fn
#undef STORE_method_set_delete_function
#define STORE_method_set_delete_function STORE_meth_set_delete_fn
#undef STORE_method_set_list_start_function
#define STORE_method_set_list_start_function STORE_meth_set_list_start_fn
#undef STORE_method_set_list_next_function
#define STORE_method_set_list_next_function STORE_meth_set_list_next_fn
#undef STORE_method_set_list_end_function
#define STORE_method_set_list_end_function STORE_meth_set_list_end_fn
#undef STORE_method_set_update_store_function
#define STORE_method_set_update_store_function STORE_meth_set_update_store_fn
#undef STORE_method_set_lock_store_function
#define STORE_method_set_lock_store_function STORE_meth_set_lock_store_fn
#undef STORE_method_set_unlock_store_function
#define STORE_method_set_unlock_store_function STORE_meth_set_unlock_store_fn
#undef STORE_method_get_initialise_function
#define STORE_method_get_initialise_function STORE_meth_get_initialise_fn
#undef STORE_method_get_cleanup_function
#define STORE_method_get_cleanup_function STORE_meth_get_cleanup_fn
#undef STORE_method_get_generate_function
#define STORE_method_get_generate_function STORE_meth_get_generate_fn
#undef STORE_method_get_modify_function
#define STORE_method_get_modify_function STORE_meth_get_modify_fn
#undef STORE_method_get_revoke_function
#define STORE_method_get_revoke_function STORE_meth_get_revoke_fn
#undef STORE_method_get_delete_function
#define STORE_method_get_delete_function STORE_meth_get_delete_fn
#undef STORE_method_get_list_start_function
#define STORE_method_get_list_start_function STORE_meth_get_list_start_fn
#undef STORE_method_get_list_next_function
#define STORE_method_get_list_next_function STORE_meth_get_list_next_fn
#undef STORE_method_get_list_end_function
#define STORE_method_get_list_end_function STORE_meth_get_list_end_fn
#undef STORE_method_get_update_store_function
#define STORE_method_get_update_store_function STORE_meth_get_update_store_fn
#undef STORE_method_get_lock_store_function
#define STORE_method_get_lock_store_function STORE_meth_get_lock_store_fn
#undef STORE_method_get_unlock_store_function
#define STORE_method_get_unlock_store_function STORE_meth_get_unlock_store_fn
/* Hack some long TS names */
#undef TS_RESP_CTX_set_status_info_cond
#define TS_RESP_CTX_set_status_info_cond TS_RESP_CTX_set_stat_info_cond
#undef TS_RESP_CTX_set_clock_precision_digits
#define TS_RESP_CTX_set_clock_precision_digits TS_RESP_CTX_set_clk_prec_digits
#undef TS_CONF_set_clock_precision_digits
#define TS_CONF_set_clock_precision_digits TS_CONF_set_clk_prec_digits
/* Hack some long CMS names */
#undef CMS_RecipientInfo_ktri_get0_algs
#define CMS_RecipientInfo_ktri_get0_algs CMS_RecipInfo_ktri_get0_algs
#undef CMS_RecipientInfo_ktri_get0_signer_id
#define CMS_RecipientInfo_ktri_get0_signer_id CMS_RecipInfo_ktri_get0_sigr_id
#undef CMS_OtherRevocationInfoFormat_it
#define CMS_OtherRevocationInfoFormat_it CMS_OtherRevocInfoFormat_it
#undef CMS_KeyAgreeRecipientIdentifier_it
#define CMS_KeyAgreeRecipientIdentifier_it CMS_KeyAgreeRecipIdentifier_it
#undef CMS_OriginatorIdentifierOrKey_it
#define CMS_OriginatorIdentifierOrKey_it CMS_OriginatorIdOrKey_it
#undef cms_SignerIdentifier_get0_signer_id
#define cms_SignerIdentifier_get0_signer_id cms_SignerId_get0_signer_id
/* Hack some long DTLS1 names */
#undef dtls1_retransmit_buffered_messages
#define dtls1_retransmit_buffered_messages dtls1_retransmit_buffered_msgs
/* Hack some long SRP names */
#undef SRP_generate_server_master_secret
#define SRP_generate_server_master_secret SRP_gen_server_master_secret
#undef SRP_generate_client_master_secret
#define SRP_generate_client_master_secret SRP_gen_client_master_secret
/* Hack some long UI names */
#undef UI_method_get_prompt_constructor
#define UI_method_get_prompt_constructor UI_method_get_prompt_constructr
#undef UI_method_set_prompt_constructor
#define UI_method_set_prompt_constructor UI_method_set_prompt_constructr
#endif /* defined OPENSSL_SYS_VMS */
/* Case insensitive linking causes problems.... */
#if defined(OPENSSL_SYS_VMS) || defined(OPENSSL_SYS_OS2)
#undef ERR_load_CRYPTO_strings
#define ERR_load_CRYPTO_strings ERR_load_CRYPTOlib_strings
#undef OCSP_crlID_new
#define OCSP_crlID_new OCSP_crlID2_new
#undef d2i_ECPARAMETERS
#define d2i_ECPARAMETERS d2i_UC_ECPARAMETERS
#undef i2d_ECPARAMETERS
#define i2d_ECPARAMETERS i2d_UC_ECPARAMETERS
#undef d2i_ECPKPARAMETERS
#define d2i_ECPKPARAMETERS d2i_UC_ECPKPARAMETERS
#undef i2d_ECPKPARAMETERS
#define i2d_ECPKPARAMETERS i2d_UC_ECPKPARAMETERS
/* These functions do not seem to exist! However, I'm paranoid...
Original command in x509v3.h:
These functions are being redefined in another directory,
and clash when the linker is case-insensitive, so let's
hide them a little, by giving them an extra 'o' at the
beginning of the name... */
#undef X509v3_cleanup_extensions
#define X509v3_cleanup_extensions oX509v3_cleanup_extensions
#undef X509v3_add_extension
#define X509v3_add_extension oX509v3_add_extension
#undef X509v3_add_netscape_extensions
#define X509v3_add_netscape_extensions oX509v3_add_netscape_extensions
#undef X509v3_add_standard_extensions
#define X509v3_add_standard_extensions oX509v3_add_standard_extensions
/* This one clashes with CMS_data_create */
#undef cms_Data_create
#define cms_Data_create priv_cms_Data_create
#endif
#endif /* ! defined HEADER_VMS_IDHACKS_H */
+31 -28
View File
@@ -54,39 +54,42 @@ static const char rcsid[]="$Id: getopt_long.c,v 1.1 2009/10/16 19:50:28 rodney E
* POSSIBILITY OF SUCH DAMAGE.
*/
#if 0
#include <err.h>
#endif
#include <errno.h>
#include <getopt.h>
#include <stdlib.h>
#include <string.h>
#ifdef _WIN32
/* Windows needs warnx(). We change the definition though:
* 1. (another) global is defined, opterrmsg, which holds the error message
* 2. errors are always printed out on stderr w/o the program name
* Note that opterrmsg always gets set no matter what opterr is set to. The
* error message will not be printed if opterr is 0 as usual.
*/
#include <stdio.h>
#include <stdarg.h>
#include <stdarg.h>
char opterrmsg[128]; /* last error message is stored here */
static void warnx(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (fmt != NULL)
_vsnprintf(opterrmsg, 128, fmt, ap);
else
opterrmsg[0]='\0';
va_end(ap);
fprintf(stderr, opterrmsg);
fprintf(stderr, "\n");
}
#ifdef _WIN32
/* Windows needs warnx(). We change the definition though:
* 1. (another) global is defined, opterrmsg, which holds the error message
* 2. errors are always printed out on stderr w/o the program name
* Note that opterrmsg always gets set no matter what opterr is set to. The
* error message will not be printed if opterr is 0 as usual.
*/
#include <stdio.h>
#include <stdarg.h>
#include <stdarg.h>
char opterrmsg[128]; /* last error message is stored here */
static void warnx(const char *fmt, ...)
{
va_list ap;
va_start(ap, fmt);
if (fmt != NULL)
_vsnprintf(opterrmsg, 128, fmt, ap);
else
opterrmsg[0]='\0';
va_end(ap);
fprintf(stderr, opterrmsg);
fprintf(stderr, "\n");
}
#endif /*_WIN32*/
#define REPLACE_GETOPT /* use this getopt as the system getopt(3) */
+1 -8
View File
@@ -1,4 +1,4 @@
AC_INIT([ccminer], [1.6.6-dev])
AC_INIT([ccminer], [1.6.4])
AC_PREREQ([2.59c])
AC_CANONICAL_SYSTEM
@@ -72,16 +72,9 @@ AC_CHECK_LIB([pthread], [pthread_create], PTHREAD_LIBS="-lpthread",
AC_CHECK_LIB([pthreadGC], [pthread_create], PTHREAD_LIBS="-lpthreadGC"
))))
AC_CHECK_LIB([z],[gzopen], [], [])
AC_CHECK_LIB([ssl],[SSL_library_init], [], [AC_MSG_ERROR([OpenSSL library required])])
AC_CHECK_LIB([crypto],[EVP_DigestFinal_ex], [], [AC_MSG_ERROR([OpenSSL library required])])
# libpci-dev (to get vendor name)
PCILIB=""; PCIFLAGS=""
AC_CHECK_LIB([pci],[pci_alloc], PCILIB="-lpci"; PCIFLAGS="-DHAVE_PCIDEV", [])
AC_SUBST(PCILIB)
AC_SUBST(PCIFLAGS)
AM_CONDITIONAL([WANT_JANSSON], [test x$request_jansson = xtrue])
AM_CONDITIONAL([HAVE_WINDOWS], [test x$have_win32 = xtrue])
AM_CONDITIONAL([ARCH_x86], [test x$have_x86 = xtrue])
+3 -6
View File
@@ -47,9 +47,6 @@
/* Define to 1 if you have the `ssl' library (-lssl). */
#define HAVE_LIBSSL 1
/* Define to 1 if you have the `z' library (-lz). */
#define HAVE_LIBZ 1
/* Define to 1 if you have the <memory.h> header file. */
#define HAVE_MEMORY_H 1
@@ -162,7 +159,7 @@
#define PACKAGE_NAME "ccminer"
/* Define to the full name and version of this package. */
#define PACKAGE_STRING "ccminer 1.6.6-dev"
#define PACKAGE_STRING "ccminer 1.6.4"
/* Define to the one symbol short name of this package. */
#define PACKAGE_TARNAME "ccminer"
@@ -171,7 +168,7 @@
#define PACKAGE_URL "http://github.com/tpruvot/ccminer"
/* Define to the version of this package. */
#define PACKAGE_VERSION "1.6.6-dev"
#define PACKAGE_VERSION "1.6.4"
/* If using the C implementation of alloca, define if you know the
direction of stack growth for your system; otherwise it will be
@@ -185,7 +182,7 @@
#define STDC_HEADERS 1
/* Version number of package */
#define VERSION "1.6.6-dev"
#define VERSION "1.6.4"
/* Define curl_free() as free() if our version of curl lacks curl_free. */
/* #undef curl_free */
+8 -35
View File
@@ -1,7 +1,6 @@
#include <stdio.h>
#include <memory.h>
#include <string.h>
#include <unistd.h>
#include <map>
#ifndef _WIN32
@@ -19,7 +18,6 @@
#endif
#include "miner.h"
#include "nvml.h"
#include "cuda_runtime.h"
@@ -62,43 +60,25 @@ void cuda_devicenames()
exit(1);
}
GPU_N = min(MAX_GPUS, GPU_N);
for (int i=0; i < GPU_N; i++)
{
char vendorname[32] = { 0 };
cudaDeviceProp props;
cudaGetDeviceProperties(&props, device_map[i]);
device_name[i] = strdup(props.name);
device_sm[i] = (props.major * 100 + props.minor * 10);
if (device_name[i]) {
free(device_name[i]);
device_name[i] = NULL;
}
#ifdef USE_WRAPNVML
if (gpu_vendor((uint8_t)props.pciBusID, vendorname) > 0 && strlen(vendorname)) {
device_name[i] = (char*) calloc(1, strlen(vendorname) + strlen(props.name) + 2);
if (!strncmp(props.name, "GeForce ", 8))
sprintf(device_name[i], "%s %s", vendorname, &props.name[8]);
else
sprintf(device_name[i], "%s %s", vendorname, props.name);
} else
#endif
device_name[i] = strdup(props.name);
}
}
void cuda_print_devices()
{
int ngpus = cuda_num_devices();
cuda_devicenames();
for (int n=0; n < ngpus; n++) {
int m = device_map[n];
cudaDeviceProp props;
cudaGetDeviceProperties(&props, m);
if (!opt_n_threads || n < opt_n_threads) {
fprintf(stderr, "GPU #%d: SM %d.%d %s\n", m, props.major, props.minor, device_name[n]);
}
if (!opt_n_threads || n < opt_n_threads)
fprintf(stderr, "GPU #%d: SM %d.%d %s\n", m, props.major, props.minor, props.name);
}
}
@@ -197,20 +177,13 @@ int cuda_gpu_clocks(struct cgpu_info *gpu)
void cuda_reset_device(int thr_id, bool *init)
{
int dev_id = device_map[thr_id];
cudaSetDevice(dev_id);
if (init != NULL) {
// with init array, its meant to be used in algo's scan code...
for (int i=0; i < MAX_GPUS; i++) {
if (device_map[i] == dev_id) {
init[i] = false;
}
for (int i=0; i < MAX_GPUS; i++) {
if (device_map[i] == dev_id) {
init[i] = false;
}
// force exit from algo's scan loops/function
restart_threads();
cudaDeviceSynchronize();
while (cudaStreamQuery(NULL) == cudaErrorNotReady)
usleep(1000);
}
restart_threads();
cudaDeviceSynchronize();
cudaDeviceReset();
}
-18
View File
@@ -13,14 +13,12 @@ __constant__ uint32_t pTarget[8]; // 32 bytes
// store MAX_GPUS device arrays of 8 nonces
static uint32_t* h_resNonces[MAX_GPUS];
static uint32_t* d_resNonces[MAX_GPUS];
static bool init_done = false;
__host__
void cuda_check_cpu_init(int thr_id, uint32_t threads)
{
CUDA_CALL_OR_RET(cudaMallocHost(&h_resNonces[thr_id], 32));
CUDA_CALL_OR_RET(cudaMalloc(&d_resNonces[thr_id], 32));
init_done = true;
}
// Target Difficulty
@@ -99,11 +97,6 @@ uint32_t cuda_check_hash(int thr_id, uint32_t threads, uint32_t startNounce, uin
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
if (!init_done) {
applog(LOG_ERR, "missing call to cuda_check_cpu_init");
return UINT32_MAX;
}
cuda_checkhash_64 <<<grid, block>>> (threads, startNounce, d_inputHash, d_resNonces[thr_id]);
cudaThreadSynchronize();
@@ -137,11 +130,6 @@ uint32_t cuda_check_hash_suppl(int thr_id, uint32_t threads, uint32_t startNounc
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
if (!init_done) {
applog(LOG_ERR, "missing call to cuda_check_cpu_init");
return 0;
}
// first element stores the count of found nonces
cudaMemset(d_resNonces[thr_id], 0, sizeof(uint32_t));
@@ -192,12 +180,6 @@ uint32_t cuda_check_hash_branch(int thr_id, uint32_t threads, uint32_t startNoun
const uint32_t threadsperblock = 256;
uint32_t result = UINT32_MAX;
if (!init_done) {
applog(LOG_ERR, "missing call to cuda_check_cpu_init");
return UINT32_MAX;
}
cudaMemset(d_resNonces[thr_id], 0xff, sizeof(uint32_t));
dim3 grid((threads + threadsperblock-1)/threadsperblock);
+50 -35
View File
@@ -148,13 +148,6 @@ __device__ __forceinline__ uint64_t cuda_swab64(uint64_t x)
(((uint64_t)(x) & 0x00000000000000ffULL) << 56)))
#endif
// swap two uint32_t without extra registers
__device__ __host__ __forceinline__ void xchg(uint32_t &x, uint32_t &y) {
x ^= y; y = x ^ y; x ^= y;
}
// for other types...
#define XCHG(x, y) { x ^= y; y = x ^ y; x ^= y; }
/*********************************************************************/
// Macros to catch CUDA errors in CUDA runtime calls
@@ -309,6 +302,7 @@ uint64_t shr_t64(uint64_t x, uint32_t n)
#endif
}
// device asm for ?
__device__ __forceinline__
uint64_t shl_t64(uint64_t x, uint32_t n)
{
@@ -323,6 +317,40 @@ uint64_t shl_t64(uint64_t x, uint32_t n)
#endif
}
// device asm 32 for pluck
__device__ __forceinline__
uint32_t andor32(uint32_t a, uint32_t b, uint32_t c) {
#ifdef __CUDA_ARCH__
uint32_t result;
asm("{ .reg .u32 m,n,o;\n\t"
"and.b32 m, %1, %2;\n\t"
" or.b32 n, %1, %2;\n\t"
"and.b32 o, n, %3;\n\t"
" or.b32 %0, m, o ;\n\t"
"}\n\t"
: "=r"(result) : "r"(a), "r"(b), "r"(c));
return result;
#else
// unused on host...
return 0;
#endif
}
__device__ __forceinline__
uint32_t xor3b(uint32_t a, uint32_t b, uint32_t c) {
#ifdef __CUDA_ARCH__
uint32_t result;
asm("{ .reg .u32 t1;\n\t"
"xor.b32 t1, %2, %3;\n\t"
"xor.b32 %0, %1, t1;\n\t"
"}"
: "=r"(result) : "r"(a) ,"r"(b),"r"(c));
return result;
#else
return a^b^c;
#endif
}
__device__ __forceinline__
uint32_t shr_t32(uint32_t x,uint32_t n) {
#ifdef __CUDA_ARCH__
@@ -452,7 +480,7 @@ uint64_t SWAPDWORDS(uint64_t value)
#endif
}
/* lyra2/bmw - uint2 vector's operators */
/* lyra2 - int2 operators */
__device__ __forceinline__
void LOHI(uint32_t &lo, uint32_t &hi, uint64_t x) {
@@ -460,41 +488,27 @@ void LOHI(uint32_t &lo, uint32_t &hi, uint64_t x) {
asm("mov.b64 {%0,%1},%2; \n\t"
: "=r"(lo), "=r"(hi) : "l"(x));
#else
lo = (uint32_t)(x);
hi = (uint32_t)(x >> 32);
lo = _LODWORD(x);
hi = _HIDWORD(x);
#endif
}
static __host__ __device__ __forceinline__ uint2 vectorize(uint64_t v) {
static __device__ __forceinline__ uint2 vectorize(uint64_t v) {
uint2 result;
#ifdef __CUDA_ARCH__
asm("mov.b64 {%0,%1},%2; \n\t"
: "=r"(result.x), "=r"(result.y) : "l"(v));
#else
result.x = (uint32_t)(v);
result.y = (uint32_t)(v >> 32);
#endif
LOHI(result.x, result.y, v);
return result;
}
static __host__ __device__ __forceinline__ uint64_t devectorize(uint2 v) {
#ifdef __CUDA_ARCH__
return MAKE_ULONGLONG(v.x, v.y);
#else
return (((uint64_t)v.y) << 32) + v.x;
#endif
}
static __device__ __forceinline__ uint64_t devectorize(uint2 v) { return MAKE_ULONGLONG(v.x, v.y); }
/**
* uint2 direct ops by c++ operator definitions
*/
static __device__ __forceinline__ uint2 operator^ (uint2 a, uint2 b) { return make_uint2(a.x ^ b.x, a.y ^ b.y); }
static __device__ __forceinline__ uint2 operator& (uint2 a, uint2 b) { return make_uint2(a.x & b.x, a.y & b.y); }
static __device__ __forceinline__ uint2 operator| (uint2 a, uint2 b) { return make_uint2(a.x | b.x, a.y | b.y); }
static __device__ __forceinline__ uint2 operator~ (uint2 a) { return make_uint2(~a.x, ~a.y); }
static __device__ __forceinline__ void operator^= (uint2 &a, uint2 b) { a = a ^ b; }
static __device__ __forceinline__ uint2 operator+ (uint2 a, uint2 b) {
static __device__ __forceinline__ uint2 operator+ (uint2 a, uint2 b)
{
#ifdef __CUDA_ARCH__
uint2 result;
asm("{\n\t"
@@ -504,14 +518,15 @@ static __device__ __forceinline__ uint2 operator+ (uint2 a, uint2 b) {
: "=r"(result.x), "=r"(result.y) : "r"(a.x), "r"(a.y), "r"(b.x), "r"(b.y));
return result;
#else
return vectorize(devectorize(a) + devectorize(b));
// incorrect but unused host equiv
return make_uint2(a.x + b.x, a.y + b.y);
#endif
}
static __device__ __forceinline__ void operator+= (uint2 &a, uint2 b) { a = a + b; }
static __device__ __forceinline__ uint2 operator- (uint2 a, uint2 b) {
#if defined(__CUDA_ARCH__) && CUDA_VERSION < 7000
static __device__ __forceinline__ uint2 operator- (uint2 a, uint2 b)
{
#ifdef __CUDA_ARCH__
uint2 result;
asm("{\n\t"
"sub.cc.u32 %0,%2,%4; \n\t"
@@ -520,10 +535,10 @@ static __device__ __forceinline__ uint2 operator- (uint2 a, uint2 b) {
: "=r"(result.x), "=r"(result.y) : "r"(a.x), "r"(a.y), "r"(b.x), "r"(b.y));
return result;
#else
return vectorize(devectorize(a) - devectorize(b));
// incorrect but unused host equiv
return make_uint2(a.x - b.x, a.y - b.y);
#endif
}
static __device__ __forceinline__ void operator-= (uint2 &a, uint2 b) { a = a - b; }
/**
* basic multiplication between 64bit no carry outside that range (ie mul.lo.b64(a*b))
-531
View File
@@ -1,531 +0,0 @@
/*
* Genoil's CUDA mining kernel for Ethereum
* based on Tim Hughes' opencl kernel.
* thanks to tpruvot,djm34,sp,cbuchner for things i took from ccminer.
*/
#define SHUFFLE_MIN_VER 9350
#include "miner.h"
#include "cuda_helper.h"
#include "ethash_cu_miner_kernel.h"
#include <stdio.h>
#include <unistd.h>
#include <memory.h>
#include <fcntl.h>
#include <sys/stat.h>
#include <sys/mman.h>
#include "vector_types.h"
#include "rotl64.cuh"
#include "keccak.cuh"
//#include "ethash_cu_miner_kernel_globals.h"
__constant__ uint32_t d_dag_size;
__constant__ uint32_t d_max_outputs;
#define ACCESSES 64
#define THREADS_PER_HASH (128 / 16) /* 8 */
#define FNV_PRIME 0x01000193
#define SWAP64(v) \
((ROTL64L(v, 8) & 0x000000FF000000FF) | \
(ROTL64L(v, 24) & 0x0000FF000000FF00) | \
(ROTL64H(v, 40) & 0x00FF000000FF0000) | \
(ROTL64H(v, 56) & 0xFF000000FF000000))
#undef SWAP64
#define SWAP64(v) cuda_swab64(v)
#define PACK64(result, lo, hi) asm("mov.b64 %0, {%1,%2};//pack64" : "=l"(result) : "r"(lo), "r"(hi));
#define UNPACK64(lo, hi, input) asm("mov.b64 {%0, %1}, %2;//unpack64" : "=r"(lo),"=r"(hi) : "l"(input));
#define copy(dst, src, count) for (uint32_t i = 0; i < count; i++) { (dst)[i] = (src)[i]; }
#define countof(x) (sizeof(x) / sizeof(x[0]))
#define fnv(x,y) ((x) * FNV_PRIME ^(y))
__device__ uint4 fnv4(uint4 a, uint4 b)
{
uint4 c;
c.x = a.x * FNV_PRIME ^ b.x;
c.y = a.y * FNV_PRIME ^ b.y;
c.z = a.z * FNV_PRIME ^ b.z;
c.w = a.w * FNV_PRIME ^ b.w;
return c;
}
__device__ uint32_t fnv_reduce(uint4 v)
{
return fnv(fnv(fnv(v.x, v.y), v.z), v.w);
}
__device__ void init_hash(hash32_t const* header, uint64_t nonce, hash64_t &init)
{
// sha3_512(header .. nonce)
uint64_t state[25] = { 0 };
//copy(state, header->uint64s, 4);
#pragma unroll
for (int i = 0; i < 4; i++) state[i] = header->uint64s[i];
state[4] = nonce;
state[5] = 0x0000000000000001;
//state[6] = 0;
//state[7] = 0;
state[8] = 0x8000000000000000;
//for (uint32_t i = 9; i < 25; i++) {
// state[i] = 0;
//}
keccak_f1600_block(state, 8);
//copy(init.uint64s, state, 8);
#pragma unroll
for (int i = 0; i < 8; i++) init.uint64s[i] = state[i];
}
__device__ uint32_t inner_loop(uint4 mix, uint32_t thread_id, uint32_t* share, hash128_t const* g_dag)
{
// share init0
if (thread_id == 0)
*share = mix.x;
uint32_t init0 = *share;
uint32_t a = 0;
do
{
bool update_share = thread_id == ((a >> 2) & (THREADS_PER_HASH-1));
//#pragma unroll 4
for (uint32_t i = 0; i < 4; i++)
{
if (update_share)
{
uint32_t m[4] = { mix.x, mix.y, mix.z, mix.w };
*share = fnv(init0 ^ (a + i), m[i]) % d_dag_size;
}
__threadfence_block();
#if __CUDA_ARCH__ >= 350
mix = fnv4(mix, __ldg(&g_dag[*share].uint4s[thread_id]));
#else
mix = fnv4(mix, g_dag[*share].uint4s[thread_id]);
#endif
}
} while ((a += 4) != ACCESSES);
return fnv_reduce(mix);
}
__device__ void final_hash(hash64_t const* init, hash32_t const* mix, hash32_t &hash)
{
uint64_t state[25] = { 0 };
// keccak_256(keccak_512(header..nonce) .. mix);
copy(state, init->uint64s, 8);
copy(state + 8, mix->uint64s, 4);
state[12] = 0x0000000000000001;
//#pragma unroll
//for (uint32_t i = 13; i < 16; i++)
//{
//state[i] = 0;
//}
state[16] = 0x8000000000000000;
//#pragma unroll
//for (uint32_t i = 17; i < 25; i++)
//{
//state[i] = 0;
//}
keccak_f1600_block(state,1);
// copy out
copy(hash.uint64s, state, 4);
}
typedef union
{
hash64_t init;
hash32_t mix;
} compute_hash_share;
#if __CUDA_ARCH__ >= SHUFFLE_MIN_VER
__device__ uint64_t compute_hash_shuffle(
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t nonce
)
{
// sha3_512(header .. nonce)
uint64_t state[25];
copy(state, g_header->uint64s, 4);
state[4] = nonce;
state[5] = 0x0000000000000001ULL;
for (uint32_t i = 6; i < 25; i++)
{
state[i] = 0;
}
state[8] = 0x8000000000000000ULL;
keccak_f1600_block(state, 8);
// Threads work together in this phase in groups of 8.
const int thread_id = threadIdx.x & (THREADS_PER_HASH - 1);
const int start_lane = threadIdx.x & ~(THREADS_PER_HASH - 1);
const int mix_idx = (thread_id & 3);
uint4 mix;
uint32_t shuffle[16];
//uint32_t * init = (uint32_t *)state;
uint32_t init[16];
UNPACK64(init[0], init[1], state[0]);
UNPACK64(init[2], init[3], state[1]);
UNPACK64(init[4], init[5], state[2]);
UNPACK64(init[6], init[7], state[3]);
UNPACK64(init[8], init[9], state[4]);
UNPACK64(init[10], init[11], state[5]);
UNPACK64(init[12], init[13], state[6]);
UNPACK64(init[14], init[15], state[7]);
for (int i = 0; i < THREADS_PER_HASH; i++)
{
// share init among threads
for (int j = 0; j < 16; j++)
shuffle[j] = __shfl(init[j], start_lane + i);
// ugly but avoids local reads/writes
if (mix_idx == 0) {
mix = make_uint4(shuffle[0], shuffle[1], shuffle[2], shuffle[3]);
}
else if (mix_idx == 1) {
mix = make_uint4(shuffle[4], shuffle[5], shuffle[6], shuffle[7]);
}
else if (mix_idx == 2) {
mix = make_uint4(shuffle[8], shuffle[9], shuffle[10], shuffle[11]);
}
else {
mix = make_uint4(shuffle[12], shuffle[13], shuffle[14], shuffle[15]);
}
uint32_t init0 = __shfl(shuffle[0], start_lane);
for (uint32_t a = 0; a < ACCESSES; a+=4)
{
int t = ((a >> 2) & (THREADS_PER_HASH - 1));
for (uint32_t b = 0; b < 4; b++)
{
if (thread_id == t)
{
shuffle[0] = fnv(init0 ^ (a + b), ((uint32_t *)&mix)[b]) % d_dag_size;;
}
shuffle[0] = __shfl(shuffle[0], start_lane + t);
mix = fnv4(mix, g_dag[shuffle[0]].uint4s[thread_id]);
}
}
uint32_t thread_mix = fnv_reduce(mix);
// update mix accross threads
for (int j = 0; j < 8; j++)
shuffle[j] = __shfl(thread_mix, start_lane + j);
if (i == thread_id) {
//move mix into state:
PACK64(state[8], shuffle[0], shuffle[1]);
PACK64(state[9], shuffle[2], shuffle[3]);
PACK64(state[10], shuffle[4], shuffle[5]);
PACK64(state[11], shuffle[6], shuffle[7]);
}
}
// keccak_256(keccak_512(header..nonce) .. mix);
state[12] = 0x0000000000000001ULL;
for (uint32_t i = 13; i < 25; i++)
{
state[i] = 0ULL;
}
state[16] = 0x8000000000000000;
keccak_f1600_block(state, 1);
return state[0];
}
#endif
__device__ void compute_hash(
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t nonce,
hash32_t &out
)
{
extern __shared__ compute_hash_share share[];
// Compute one init hash per work item.
hash64_t init;
init_hash(g_header, nonce, init);
// Threads work together in this phase in groups of 8.
uint32_t const thread_id = threadIdx.x & (THREADS_PER_HASH-1);
uint32_t const hash_id = threadIdx.x >> 3;
hash32_t mix;
for (int i = 0; i < THREADS_PER_HASH; i++)
{
// share init with other threads
if (i == thread_id)
share[hash_id].init = init;
uint4 thread_init = share[hash_id].init.uint4s[thread_id & 3];
uint32_t thread_mix = inner_loop(thread_init, thread_id, share[hash_id].mix.uint32s, g_dag);
share[hash_id].mix.uint32s[thread_id] = thread_mix;
if (i == thread_id)
mix = share[hash_id].mix;
}
final_hash(&init, &mix, out);
}
__global__ void
__launch_bounds__(128, 7)
ethash_search(
uint32_t* g_output,
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t start_nonce,
uint64_t target
)
{
uint32_t const gid = blockIdx.x * blockDim.x + threadIdx.x;
#if 0 // _CUDA_ARCH__ >= SHUFFLE_MIN_VER
uint64_t hash = compute_hash_shuffle(g_header, g_dag, start_nonce + gid);
if (SWAP64(hash) < target)
{
atomicInc(g_output, d_max_outputs);
g_output[g_output[0]] = gid;
}
#else
hash32_t hash;
compute_hash(g_header, g_dag, start_nonce + gid, hash);
if (SWAP64(hash.uint64s[0]) <= target)
{
atomicInc(g_output,d_max_outputs);
g_output[g_output[0]] = gid;
}
#endif
}
void run_ethash_hash(
hash32_t* g_hashes,
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t start_nonce
)
{
}
void run_ethash_search(
uint32_t blocks,
uint32_t threads,
cudaStream_t stream,
uint32_t* g_output,
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t start_nonce,
uint64_t target
)
{
#if __CUDA_ARCH__ >= SHUFFLE_MIN_VER
ethash_search <<<blocks, threads, 0, stream>>>(g_output, g_header, g_dag, start_nonce, target);
#else
ethash_search <<<blocks, threads, (sizeof(compute_hash_share) * threads) / THREADS_PER_HASH, stream>>>(g_output, g_header, g_dag, start_nonce, target);
#endif
}
cudaError set_constants(
uint32_t * dag_size,
uint32_t * max_outputs
)
{
cudaError result;
result = cudaMemcpyToSymbol(d_dag_size, dag_size, sizeof(uint32_t));
result = cudaMemcpyToSymbol(d_max_outputs, max_outputs, sizeof(uint32_t));
return result;
}
static unsigned m_search_batch_size = 262144;
static unsigned m_workgroup_size = 64;
static __thread hash128_t * m_dag_ptr;
static __thread hash32_t * m_header;
//static void* m_hash_buf[MAX_GPUS] = { 0 };
static uint32_t* m_search_buf[MAX_GPUS] = { 0 };
static cudaStream_t m_streams[MAX_GPUS] = { 0 };
static uint64_t dagSize = 0;
static int fdDAG = -1;
static uint8_t const *dag = NULL;
uint32_t const c_zero = 0;
#define ETHASH_MIX_BYTES 128
#define MIX_WORDS (ETHASH_MIX_BYTES/4)
uint8_t* alloc_dag_file(uint32_t rev, uint32_t* seedhash)
{
uint8_t* buffer = NULL;
char fn[256];
struct stat stat_buf;
int rc;
sprintf(fn, "%s/.ethash/full-R%u-%08x%08x", getenv("HOME"), rev, seedhash[7], seedhash[6]);
rc = stat(fn, &stat_buf);
fdDAG = open(fn, O_RDONLY);
if (rc != 0) {
applog(LOG_ERR, "Unable to open DAG file %s (%d)!", fn, rc);
return NULL;
}
dagSize = stat_buf.st_size - 8;
#ifdef __linux__
buffer = (uint8_t*) (mmap(NULL, dagSize, PROT_READ, MAP_FILE | MAP_SHARED | MAP_POPULATE , fdDAG, 0));
#else
buffer = (uint8_t*) (mmap(NULL, dagSize, PROT_READ, MAP_FILE | MAP_SHARED), fdDAG, 0));
#endif
if (buffer == MAP_FAILED) {
applog(LOG_ERR, "Unable to map DAG file %s!", fn);
return NULL;
}
if (buffer[0] != 0xfe) {
applog(LOG_ERR, "DAG file header is bad %s!", fn);
close(fdDAG);
return NULL;
}
return (buffer+8);
}
void close_dag_file(uint8_t const *dag)
{
uint8_t* buffer = dag ? (uint8_t*) dag - 8 : NULL;
munmap(buffer, dagSize);
dag = NULL;
close(fdDAG);
fdDAG = -1;
}
static bool init[MAX_GPUS] = { 0 };
int scanhash_ether(int thr_id, struct work *work, uint64_t max_nonce, unsigned long *hashes_done)
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t *pseed = &work->data[8];
const uint64_t first_nonce = (((uint64_t)pdata[20]) << 32) + pdata[19];
uint64_t nonce = first_nonce;
uint32_t throughput = device_intensity(thr_id, __func__, 1U << 18);
throughput = min(throughput, max_nonce - first_nonce);
if (!dagSize) {
dag = alloc_dag_file(23, pseed);
applog(LOG_NOTICE, "DAG size=%ld", (long int) dagSize);
}
//0000.0000 0050.4c39c 047378ad262a995be73cc504643443728e1ca415bae9208
uint64_t target64 = ptarget[7];
target64 = (target64 << 32) + ptarget[6];
if (!init[thr_id])
{
uint32_t dag_size = (unsigned)(dagSize / ETHASH_MIX_BYTES);
uint32_t max_outputs=63;
cudaSetDevice(device_map[thr_id]);
// create buffer for header256
cudaMalloc(&m_header, 32);
// create mining buffer
cudaMalloc(&m_search_buf[thr_id], 8 * sizeof(uint32_t));
//cudaMalloc(&m_hash_buf[thr_id], c_hash_batch_size * 32);
//cudaStreamCreate(&m_streams[thr_id]);
m_streams[thr_id] = 0;
// create buffer for dag
CUDA_SAFE_CALL(cudaMalloc(&m_dag_ptr, dagSize));
CUDA_SAFE_CALL(cudaMemcpy(m_dag_ptr, dag, dagSize, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(set_constants(&dag_size, &max_outputs));
init[thr_id] = true;
}
else if (dag) {
close_dag_file(dag);
}
CUDA_SAFE_CALL(cudaMemcpy(m_header, work->data, 32, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMemcpy(m_search_buf[thr_id], &c_zero, 4, cudaMemcpyHostToDevice));
uint32_t results[8] = { 0 };
do {
results[0] = 0;
CUDA_SAFE_CALL(cudaMemset(m_search_buf[thr_id], 0, sizeof(uint32_t)));
run_ethash_search(m_search_batch_size / m_workgroup_size, m_workgroup_size, m_streams[thr_id], m_search_buf[thr_id], m_header, m_dag_ptr, nonce, target64);
CUDA_SAFE_CALL(cudaThreadSynchronize());
CUDA_SAFE_CALL(cudaMemcpyAsync(results, m_search_buf[thr_id], 2*sizeof(uint32_t), cudaMemcpyDeviceToHost, m_streams[thr_id]));
unsigned num_found = results[0];
if (num_found) {
nonce = nonce + results[1];
pdata[19] = (uint32_t) nonce;
pdata[20] = (uint32_t) (nonce >> 32);
return 1;
}
nonce += throughput;
} while (nonce < max_nonce && !work_restart[thr_id].restart);
*hashes_done = nonce - first_nonce + 1;
return 0;
}
-59
View File
@@ -1,59 +0,0 @@
#ifndef _ETHASH_CU_MINER_KERNEL_H_
#define _ETHASH_CU_MINER_KERNEL_H_
#include <stdint.h>
typedef union
{
uint64_t uint64s[16 / sizeof(uint64_t)];
uint32_t uint32s[16 / sizeof(uint32_t)];
} hash16_t;
typedef union
{
uint32_t uint32s[32 / sizeof(uint32_t)];
uint64_t uint64s[32 / sizeof(uint64_t)];
} hash32_t;
typedef union
{
uint32_t uint32s[64 / sizeof(uint32_t)];
uint64_t uint64s[64 / sizeof(uint64_t)];
uint4 uint4s[64 / sizeof(uint4)];
} hash64_t;
typedef union
{
uint32_t uint32s[128 / sizeof(uint32_t)];
uint4 uint4s[128 / sizeof(uint4)];
} hash128_t;
//typedef uint32_t hash128_t;
cudaError set_constants(
uint32_t * dag_size,
uint32_t * max_outputs
);
void run_ethash_hash(
hash32_t* g_hashes,
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t start_nonce
);
void run_ethash_search(
uint32_t search_batch_size,
uint32_t workgroup_size,
cudaStream_t stream,
uint32_t* g_output,
hash32_t const* g_header,
hash128_t const* g_dag,
uint64_t start_nonce,
uint64_t target
);
#endif
-510
View File
@@ -1,510 +0,0 @@
#include "cpuminer-config.h"
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <ctype.h>
#include <stdarg.h>
#include <unistd.h>
//#include <jansson.h>
#include <curl/curl.h>
#ifdef WIN32
#include <winsock2.h>
#include <mstcpip.h>
#else
#include <errno.h>
#include <sys/socket.h>
#include <netinet/in.h>
#include <netinet/tcp.h>
#endif
#include "miner.h"
struct data_buffer {
void *buf;
size_t len;
};
struct upload_buffer {
const void *buf;
size_t len;
size_t pos;
};
struct header_info {
char *lp_path;
char *reason;
char *stratum_url;
};
static void databuf_free(struct data_buffer *db)
{
if (!db)
return;
free(db->buf);
memset(db, 0, sizeof(*db));
}
static size_t all_data_cb(const void *ptr, size_t size, size_t nmemb,
void *user_data)
{
struct data_buffer *db = (struct data_buffer *)user_data;
size_t len = size * nmemb;
size_t oldlen, newlen;
void *newmem;
static const uchar zero = 0;
oldlen = db->len;
newlen = oldlen + len;
newmem = realloc(db->buf, newlen + 1);
if (!newmem)
return 0;
db->buf = newmem;
db->len = newlen;
memcpy((char*)db->buf + oldlen, ptr, len);
memcpy((char*)db->buf + newlen, &zero, 1); /* null terminate */
return len;
}
static size_t upload_data_cb(void *ptr, size_t size, size_t nmemb,
void *user_data)
{
struct upload_buffer *ub = (struct upload_buffer *)user_data;
unsigned int len = (unsigned int)(size * nmemb);
if (len > ub->len - ub->pos)
len = (unsigned int)(ub->len - ub->pos);
if (len) {
memcpy(ptr, (char*)ub->buf + ub->pos, len);
ub->pos += len;
}
return len;
}
#if LIBCURL_VERSION_NUM >= 0x071200
static int seek_data_cb(void *user_data, curl_off_t offset, int origin)
{
struct upload_buffer *ub = (struct upload_buffer *)user_data;
switch (origin) {
case SEEK_SET:
ub->pos = (size_t)offset;
break;
case SEEK_CUR:
ub->pos += (size_t)offset;
break;
case SEEK_END:
ub->pos = ub->len + (size_t)offset;
break;
default:
return 1; /* CURL_SEEKFUNC_FAIL */
}
return 0; /* CURL_SEEKFUNC_OK */
}
#endif
static size_t resp_hdr_cb(void *ptr, size_t size, size_t nmemb, void *user_data)
{
struct header_info *hi = (struct header_info *)user_data;
size_t remlen, slen, ptrlen = size * nmemb;
char *rem, *val = NULL, *key = NULL;
void *tmp;
val = (char*)calloc(1, ptrlen);
key = (char*)calloc(1, ptrlen);
if (!key || !val)
goto out;
tmp = memchr(ptr, ':', ptrlen);
if (!tmp || (tmp == ptr)) /* skip empty keys / blanks */
goto out;
slen = (size_t)((char*)tmp - (char*)ptr);
if ((slen + 1) == ptrlen) /* skip key w/ no value */
goto out;
memcpy(key, ptr, slen); /* store & nul term key */
key[slen] = 0;
rem = (char*)ptr + slen + 1; /* trim value's leading whitespace */
remlen = ptrlen - slen - 1;
while ((remlen > 0) && (isspace(*rem))) {
remlen--;
rem++;
}
memcpy(val, rem, remlen); /* store value, trim trailing ws */
val[remlen] = 0;
while ((*val) && (isspace(val[strlen(val) - 1]))) {
val[strlen(val) - 1] = 0;
}
if (!*val) /* skip blank value */
goto out;
if (!strcasecmp("X-Long-Polling", key)) {
hi->lp_path = val; /* X-Mining-Extensions: longpoll */
val = NULL;
}
if (!strcasecmp("X-Reject-Reason", key)) {
hi->reason = val; /* X-Mining-Extensions: reject-reason */
//applog(LOG_WARNING, "%s:%s", key, val);
val = NULL;
}
if (!strcasecmp("X-Stratum", key)) {
hi->stratum_url = val; /* steal memory reference */
val = NULL;
}
if (!strcasecmp("X-Nonce-Range", key)) {
/* todo when available: X-Mining-Extensions: noncerange */
}
out:
free(key);
free(val);
return ptrlen;
}
#if LIBCURL_VERSION_NUM >= 0x070f06
static int sockopt_keepalive_cb(void *userdata, curl_socket_t fd,
curlsocktype purpose)
{
int keepalive = 1;
int tcp_keepcnt = 3;
int tcp_keepidle = 50;
int tcp_keepintvl = 50;
#ifdef WIN32
DWORD outputBytes;
#endif
#ifndef WIN32
if (unlikely(setsockopt(fd, SOL_SOCKET, SO_KEEPALIVE, &keepalive,
sizeof(keepalive))))
return 1;
#ifdef __linux
if (unlikely(setsockopt(fd, SOL_TCP, TCP_KEEPCNT,
&tcp_keepcnt, sizeof(tcp_keepcnt))))
return 1;
if (unlikely(setsockopt(fd, SOL_TCP, TCP_KEEPIDLE,
&tcp_keepidle, sizeof(tcp_keepidle))))
return 1;
if (unlikely(setsockopt(fd, SOL_TCP, TCP_KEEPINTVL,
&tcp_keepintvl, sizeof(tcp_keepintvl))))
return 1;
#endif /* __linux */
#ifdef __APPLE_CC__
if (unlikely(setsockopt(fd, IPPROTO_TCP, TCP_KEEPALIVE,
&tcp_keepintvl, sizeof(tcp_keepintvl))))
return 1;
#endif /* __APPLE_CC__ */
#else /* WIN32 */
struct tcp_keepalive vals;
vals.onoff = 1;
vals.keepalivetime = tcp_keepidle * 1000;
vals.keepaliveinterval = tcp_keepintvl * 1000;
if (unlikely(WSAIoctl(fd, SIO_KEEPALIVE_VALS, &vals, sizeof(vals),
NULL, 0, &outputBytes, NULL, NULL)))
return 1;
#endif /* WIN32 */
return 0;
}
#endif
static json_t *ether_json_rpc(CURL *curl, const char *url, const char *rpc_req, int *curl_err)
{
char len_hdr[64], hashrate_hdr[64];
char curl_err_str[CURL_ERROR_SIZE] = { 0 };
struct data_buffer all_data = { 0 };
struct upload_buffer upload_data;
struct curl_slist *headers = NULL;
struct header_info hi = { 0 };
char *httpdata;
json_t *val, *err_val, *res_val;
json_error_t err;
int rc;
long timeout = opt_timeout;
bool keepalive = false;
/* it is assumed that 'curl' is freshly [re]initialized at this pt */
if (opt_protocol)
curl_easy_setopt(curl, CURLOPT_VERBOSE, 1);
curl_easy_setopt(curl, CURLOPT_URL, url);
if (opt_cert)
curl_easy_setopt(curl, CURLOPT_CAINFO, opt_cert);
curl_easy_setopt(curl, CURLOPT_ENCODING, "");
curl_easy_setopt(curl, CURLOPT_FAILONERROR, 0);
curl_easy_setopt(curl, CURLOPT_NOSIGNAL, 1);
curl_easy_setopt(curl, CURLOPT_TCP_NODELAY, 1);
curl_easy_setopt(curl, CURLOPT_WRITEFUNCTION, all_data_cb);
curl_easy_setopt(curl, CURLOPT_WRITEDATA, &all_data);
curl_easy_setopt(curl, CURLOPT_READFUNCTION, upload_data_cb);
curl_easy_setopt(curl, CURLOPT_READDATA, &upload_data);
#if LIBCURL_VERSION_NUM >= 0x071200
curl_easy_setopt(curl, CURLOPT_SEEKFUNCTION, &seek_data_cb);
curl_easy_setopt(curl, CURLOPT_SEEKDATA, &upload_data);
#endif
curl_easy_setopt(curl, CURLOPT_ERRORBUFFER, curl_err_str);
curl_easy_setopt(curl, CURLOPT_FOLLOWLOCATION, 1);
curl_easy_setopt(curl, CURLOPT_TIMEOUT, timeout);
curl_easy_setopt(curl, CURLOPT_HEADERFUNCTION, resp_hdr_cb);
curl_easy_setopt(curl, CURLOPT_HEADERDATA, &hi);
if (opt_proxy) {
curl_easy_setopt(curl, CURLOPT_PROXY, opt_proxy);
curl_easy_setopt(curl, CURLOPT_PROXYTYPE, opt_proxy_type);
}
#if 0
if (userpass) {
curl_easy_setopt(curl, CURLOPT_USERPWD, userpass);
curl_easy_setopt(curl, CURLOPT_HTTPAUTH, CURLAUTH_BASIC);
}
#endif
#if LIBCURL_VERSION_NUM >= 0x070f06
if (keepalive)
curl_easy_setopt(curl, CURLOPT_SOCKOPTFUNCTION, sockopt_keepalive_cb);
#endif
curl_easy_setopt(curl, CURLOPT_POST, 1);
if (opt_protocol)
applog(LOG_DEBUG, "JSON protocol request:\n%s", rpc_req);
upload_data.buf = rpc_req;
upload_data.len = strlen(rpc_req);
upload_data.pos = 0;
sprintf(len_hdr, "Content-Length: %lu", (unsigned long) upload_data.len);
sprintf(hashrate_hdr, "X-Mining-Hashrate: %llu", (unsigned long long) global_hashrate);
headers = curl_slist_append(headers, "Content-Type: application/json");
headers = curl_slist_append(headers, len_hdr);
headers = curl_slist_append(headers, "User-Agent: " USER_AGENT);
headers = curl_slist_append(headers, "X-Mining-Extensions: longpoll noncerange reject-reason");
headers = curl_slist_append(headers, hashrate_hdr);
headers = curl_slist_append(headers, "Accept:"); /* disable Accept hdr*/
headers = curl_slist_append(headers, "Expect:"); /* disable Expect hdr*/
curl_easy_setopt(curl, CURLOPT_HTTPHEADER, headers);
rc = curl_easy_perform(curl);
if (curl_err != NULL)
*curl_err = rc;
if (rc) {
if (rc != CURLE_OPERATION_TIMEDOUT) {
applog(LOG_ERR, "HTTP request failed: %s", curl_err_str);
goto err_out;
}
}
if (!all_data.buf || !all_data.len) {
applog(LOG_ERR, "Empty data received in json_rpc_call.");
goto err_out;
}
httpdata = (char*) all_data.buf;
if (*httpdata != '{' && *httpdata != '[') {
long errcode = 0;
CURLcode c = curl_easy_getinfo(curl, CURLINFO_RESPONSE_CODE, &errcode);
if (c == CURLE_OK && errcode == 401) {
applog(LOG_ERR, "You are not authorized, check your login and password.");
goto err_out;
}
}
val = JSON_LOADS(httpdata, &err);
if (!val) {
applog(LOG_ERR, "JSON decode failed(%d): %s", err.line, err.text);
if (opt_protocol)
applog(LOG_DEBUG, "%s", httpdata);
goto err_out;
}
if (opt_protocol) {
char *s = json_dumps(val, JSON_INDENT(3));
applog(LOG_DEBUG, "JSON protocol response:\n%s\n", s);
free(s);
}
/* JSON-RPC valid response returns a non-null 'result',
* and a null 'error'. */
res_val = json_object_get(val, "result");
err_val = json_object_get(val, "error");
if (!res_val || json_is_null(res_val) ||
(err_val && !json_is_null(err_val))) {
char *s = NULL;
if (err_val) {
s = json_dumps(err_val, 0);
json_t *msg = json_object_get(err_val, "message");
json_t *err_code = json_object_get(err_val, "code");
if (curl_err && json_integer_value(err_code))
*curl_err = (int) json_integer_value(err_code);
if (json_is_string(msg)) {
free(s);
s = strdup(json_string_value(msg));
if (have_longpoll && s && !strcmp(s, "method not getwork")) {
json_decref(err_val);
free(s);
goto err_out;
}
}
json_decref(err_val);
}
else
s = strdup("(unknown reason)");
if (!curl_err || opt_debug)
applog(LOG_ERR, "JSON-RPC call failed: %s", s);
free(s);
goto err_out;
}
if (hi.reason)
json_object_set_new(val, "reject-reason", json_string(hi.reason));
databuf_free(&all_data);
curl_slist_free_all(headers);
curl_easy_reset(curl);
return val;
err_out:
free(hi.lp_path);
free(hi.reason);
free(hi.stratum_url);
databuf_free(&all_data);
curl_slist_free_all(headers);
curl_easy_reset(curl);
return NULL;
}
// Made for len 32
bool hex2uint(uint32_t *target, const char *hexstr, int len)
{
char buf[32];
char pad[80];
char hex_byte[3];
char *ep;
char *p = buf;
char *hex = pad;
hex_byte[2] = '\0';
// pad, we can receive non padded numbers (0x44 => 0x0044)
size_t hexlen = strlen(hexstr);
if (hexlen < len*2)
memset(hex, '0', len*2 - hexlen);
sprintf(&hex[len*2 - hexlen],"%s", hexstr);
while (*hex && len) {
if (!hex[1]) {
applog(LOG_ERR, "hex2uint str truncated");
return false;
}
hex_byte[0] = hex[0];
hex_byte[1] = hex[1];
*p = (uchar) strtol(hex_byte, &ep, 16);
if (*ep) {
applog(LOG_ERR, "hex2uint failed on '%s'", hex_byte);
return false;
}
p++;
hex += 2;
len--;
if (len % 8 == 0) {
//..
}
}
for (int i = 0; i < 8; i++) {
be32enc(&target[7 - i], ((uint32_t*)buf)[i]);
}
return (len == 0 && *hexstr == 0) ? true : false;
}
void uint2hex(char *out, uint32_t *in, int len)
{
if (out) {
for (int i = 0; i < len/sizeof(uint32_t); i++)
sprintf(out + (i * 8), "%08x", in[7-i]);
}
}
bool ether_getwork(CURL *curl, struct pool_infos *pool, struct work *work)
{
char req[512];
bool ret = false;
sprintf(req, "{ \"jsonrpc\": \"2.0\", \"method\": \"eth_getWork\", \"params\": [], \"id\": 4 }\r\n");
json_t *result = ether_json_rpc(curl, pool->url, req, NULL);
if (result)
ret = true;
json_t *arr = json_object_get(result, "result");
if (!json_is_array(arr)) {
return false;
}
const char *header = json_string_value(json_array_get(arr, 0));
const char *seed = json_string_value(json_array_get(arr, 1));
const char *target = json_string_value(json_array_get(arr, 2));
//0x2b800e3ffcafdf7ffdd5b2ea1c0ed79c534e92b3459331eae61597943d0743bc
//0x356e5a2cc1eba076e650ac7473fccc37952b46bc2e419a200cec0c451dce2336
//0x000000000050ec7dea051867e987e3eed0066a128f161b5776a8f5da326b06ec
applog(LOG_NOTICE, "%s %s %s", header, seed, target);
hex2uint(work->data, &header[2], 32);
hex2uint(&work->data[8], &seed[2], 32);
hex2uint(work->target, &target[2], 32);
//applog(LOG_BLUE, "%08x %08x", work->target[7], work->target[6]);
return ret;
}
bool ether_submitwork(CURL *curl, struct pool_infos *pool, struct work *work)
{
char req[512];
bool ret = false;
char noncehex[32]={0}, header[80]={0}, mixhash[80]={0};
sprintf(noncehex, "0x%08x%08x", work->data[20], work->data[19]);
uint2hex(header, work->data, 32);
uint2hex(mixhash, &work->data[24], 32);
applog(LOG_NOTICE, "%s mixhash:", __func__);
applog_hash((uchar*)&work->data[24]);
sprintf(req, "{ \"jsonrpc\": \"2.0\", \"method\": \"eth_submitWork\","
"\"params\": [\"%s\", \"0x%s\", \"0x%s\"], \"id\": 4 }\r\n",
noncehex, header, mixhash
);
json_t *result = ether_json_rpc(curl, pool->url, req, NULL);
if (result)
ret = true;
json_t *res = json_object_get(result, "result");
if (!json_is_true(res)) {
applog(LOG_ERR, "not true! %s", json_dumps(result, JSON_INDENT(3)));
exit(0);
return false;
}
return ret;
}
-83
View File
@@ -1,83 +0,0 @@
__device__ __constant__ uint64_t const keccak_round_constants[24] = {
0x0000000000000001ULL, 0x0000000000008082ULL, 0x800000000000808AULL,
0x8000000080008000ULL, 0x000000000000808BULL, 0x0000000080000001ULL,
0x8000000080008081ULL, 0x8000000000008009ULL, 0x000000000000008AULL,
0x0000000000000088ULL, 0x0000000080008009ULL, 0x000000008000000AULL,
0x000000008000808BULL, 0x800000000000008BULL, 0x8000000000008089ULL,
0x8000000000008003ULL, 0x8000000000008002ULL, 0x8000000000000080ULL,
0x000000000000800AULL, 0x800000008000000AULL, 0x8000000080008081ULL,
0x8000000000008080ULL, 0x0000000080000001ULL, 0x8000000080008008ULL
};
__device__ static void keccak_f1600_block(uint64_t* s, uint32_t out_size)//, uint32_t in_size, uint32_t out_size)
{
uint64_t t[5], u, v;
for (size_t i = 0; i < 24; i++) {
/* theta: c = a[0,i] ^ a[1,i] ^ .. a[4,i] */
t[0] = s[0] ^ s[5] ^ s[10] ^ s[15] ^ s[20];
t[1] = s[1] ^ s[6] ^ s[11] ^ s[16] ^ s[21];
t[2] = s[2] ^ s[7] ^ s[12] ^ s[17] ^ s[22];
t[3] = s[3] ^ s[8] ^ s[13] ^ s[18] ^ s[23];
t[4] = s[4] ^ s[9] ^ s[14] ^ s[19] ^ s[24];
/* theta: d[i] = c[i+4] ^ rotl(c[i+1],1) */
/* theta: a[0,i], a[1,i], .. a[4,i] ^= d[i] */
u = t[4] ^ ROTL64L(t[1], 1);
s[0] ^= u; s[5] ^= u; s[10] ^= u; s[15] ^= u; s[20] ^= u;
u = t[0] ^ ROTL64L(t[2], 1);
s[1] ^= u; s[6] ^= u; s[11] ^= u; s[16] ^= u; s[21] ^= u;
u = t[1] ^ ROTL64L(t[3], 1);
s[2] ^= u; s[7] ^= u; s[12] ^= u; s[17] ^= u; s[22] ^= u;
u = t[2] ^ ROTL64L(t[4], 1);
s[3] ^= u; s[8] ^= u; s[13] ^= u; s[18] ^= u; s[23] ^= u;
u = t[3] ^ ROTL64L(t[0], 1);
s[4] ^= u; s[9] ^= u; s[14] ^= u; s[19] ^= u; s[24] ^= u;
/* rho pi: b[..] = rotl(a[..], ..) */
u = s[1];
s[1] = ROTL64H(s[6], 44);
s[6] = ROTL64L(s[9], 20);
s[9] = ROTL64H(s[22], 61);
s[22] = ROTL64H(s[14], 39);
s[14] = ROTL64L(s[20], 18);
s[20] = ROTL64H(s[2], 62);
s[2] = ROTL64H(s[12], 43);
s[12] = ROTL64L(s[13], 25);
s[13] = ROTL64L(s[19], 8);
s[19] = ROTL64H(s[23], 56);
s[23] = ROTL64H(s[15], 41);
s[15] = ROTL64L(s[4], 27);
s[4] = ROTL64L(s[24], 14);
s[24] = ROTL64L(s[21], 2);
s[21] = ROTL64H(s[8], 55);
s[8] = ROTL64H(s[16], 45);
s[16] = ROTL64H(s[5], 36);
s[5] = ROTL64L(s[3], 28);
s[3] = ROTL64L(s[18], 21);
s[18] = ROTL64L(s[17], 15);
s[17] = ROTL64L(s[11], 10);
s[11] = ROTL64L(s[7], 6);
s[7] = ROTL64L(s[10], 3);
s[10] = ROTL64L(u, 1);
// squeeze this in here
/* chi: a[i,j] ^= ~b[i,j+1] & b[i,j+2] */
u = s[0]; v = s[1]; s[0] ^= (~v) & s[2];
/* iota: a[0,0] ^= round constant */
s[0] ^= keccak_round_constants[i];
if (i == 23 && out_size == 1) return;
// continue chi
s[1] ^= (~s[2]) & s[3]; s[2] ^= (~s[3]) & s[4]; s[3] ^= (~s[4]) & u; s[4] ^= (~u) & v;
u = s[5]; v = s[6]; s[5] ^= (~v) & s[7]; s[6] ^= (~s[7]) & s[8]; s[7] ^= (~s[8]) & s[9];
if (i == 23) return;
s[8] ^= (~s[9]) & u; s[9] ^= (~u) & v;
u = s[10]; v = s[11]; s[10] ^= (~v) & s[12]; s[11] ^= (~s[12]) & s[13]; s[12] ^= (~s[13]) & s[14]; s[13] ^= (~s[14]) & u; s[14] ^= (~u) & v;
u = s[15]; v = s[16]; s[15] ^= (~v) & s[17]; s[16] ^= (~s[17]) & s[18]; s[17] ^= (~s[18]) & s[19]; s[18] ^= (~s[19]) & u; s[19] ^= (~u) & v;
u = s[20]; v = s[21]; s[20] ^= (~v) & s[22]; s[21] ^= (~s[22]) & s[23]; s[22] ^= (~s[23]) & s[24]; s[23] ^= (~s[24]) & u; s[24] ^= (~u) & v;
}
}
-64
View File
@@ -1,64 +0,0 @@
#include <cuda.h>
#include <cuda_runtime.h>
#include <stdint.h>
// 64-bit ROTATE LEFT
#if __CUDA_ARCH__ >= 320
__device__ uint64_t ROTL64H(const uint64_t x, const int offset)
{
uint64_t res;
asm("{\n\t"
".reg .u32 tl,th,vl,vh;\n\t"
"mov.b64 {tl,th}, %1;\n\t"
"shf.l.wrap.b32 vl, tl, th, %2;\n\t"
"shf.l.wrap.b32 vh, th, tl, %2;\n\t"
"mov.b64 %0, {vl,vh};\n\t"
"}"
: "=l"(res) : "l"(x), "r"(offset)
);
return res;
}
__device__ uint64_t ROTL64L(const uint64_t x, const int offset)
{
uint64_t res;
asm("{\n\t"
".reg .u32 tl,th,vl,vh;\n\t"
"mov.b64 {tl,th}, %1;\n\t"
"shf.l.wrap.b32 vl, tl, th, %2;\n\t"
"shf.l.wrap.b32 vh, th, tl, %2;\n\t"
"mov.b64 %0, {vh,vl};\n\t"
"}"
: "=l"(res) : "l"(x), "r"(offset)
);
return res;
}
#elif __CUDA_ARCH__ >= 120
#define ROTL64H(x, n) ROTL64(x,n)
#define ROTL64L(x, n) ROTL64(x,n)
__device__ __forceinline__
uint64_t ROTL64(const uint64_t x, const int offset)
{
uint64_t result;
asm("{\n\t"
".reg .b64 lhs;\n\t"
".reg .u32 roff;\n\t"
"shl.b64 lhs, %1, %2;\n\t"
"sub.u32 roff, 64, %2;\n\t"
"shr.b64 %0, %1, roff;\n\t"
"add.u64 %0, lhs, %0;\n\t"
"}\n"
: "=l"(result) : "l"(x), "r"(offset));
return result;
}
#else
#define ROTL64H(x, n) ROTL64(x,n)
#define ROTL64L(x, n) ROTL64(x,n)
/* host */
#define ROTL64(x, n) (((x) << (n)) | ((x) >> (64 - (n))))
#endif
+1
View File
@@ -2,6 +2,7 @@
#include <stdint.h>
#include <cuda_runtime.h>
#include "uint256.h"
#include "sph/sph_fugue.h"
#include "miner.h"
+1
View File
@@ -3,6 +3,7 @@
#include <cuda_runtime.h>
#include <openssl/sha.h>
#include "uint256.h"
#include "sph/sph_groestl.h"
#include "cuda_groestlcoin.h"
-3
View File
@@ -89,9 +89,6 @@ static void round_lyra(uint2 *s)
__device__ __forceinline__
void reduceDuplexRowSetup(const int rowIn, const int rowInOut, const int rowOut, uint2 state[16], uint2 Matrix[96][8])
{
#if __CUDA_ARCH__ > 500
#pragma unroll
#endif
for (int i = 0; i < 8; i++)
{
#pragma unroll
+10 -59
View File
@@ -258,8 +258,6 @@ void sha256d(unsigned char *hash, const unsigned char *data, int len);
#define HAVE_SHA256_4WAY 0
#define HAVE_SHA256_8WAY 0
struct work;
extern int scanhash_sha256d(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce, unsigned long *hashes_done);
@@ -267,7 +265,7 @@ extern int scanhash_deep(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_luffa(int thr_id, uint32_t *pdata,
extern int scanhash_doom(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
@@ -307,10 +305,6 @@ extern int scanhash_blake256(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done, int8_t blakerounds);
extern int scanhash_c11(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_fresh(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
@@ -330,6 +324,10 @@ extern int scanhash_pentablake(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_pluck(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_qubit(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
@@ -374,16 +372,10 @@ extern int scanhash_x17(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_whirlpoolx(int thr_id, uint32_t *pdata,
extern int scanhash_zr5(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_zr5(int thr_id, struct work *work, uint32_t max_nonce,
unsigned long *hashes_done);
extern int scanhash_ether(int thr_id, struct work *work, uint64_t max_nonce,
unsigned long *hashes_done);
/* api related */
void *api_thread(void *userdata);
void api_set_throughput(int thr_id, uint32_t throughput);
@@ -404,7 +396,7 @@ struct cgpu_info {
int gpu_clock;
int gpu_memclock;
size_t gpu_mem;
uint32_t gpu_power;
uint32_t gpu_usage;
double gpu_vddc;
int16_t gpu_pstate;
int16_t gpu_bus;
@@ -476,18 +468,6 @@ struct work_restart {
char padding[128 - sizeof(uint32_t)];
};
#ifdef HAVE_GETOPT_LONG
#include <getopt.h>
#else
struct option {
const char *name;
int has_arg;
int *flag;
int val;
};
#endif
extern int options_count();
extern bool opt_benchmark;
extern bool opt_debug;
extern bool opt_quiet;
@@ -505,13 +485,11 @@ extern char *opt_proxy;
extern long opt_proxy_type;
extern bool use_syslog;
extern bool use_colors;
extern int use_pok;
extern pthread_mutex_t applog_lock;
extern struct thr_info *thr_info;
extern int longpoll_thr_id;
extern int stratum_thr_id;
extern int api_thr_id;
extern volatile bool abort_flag;
extern struct work_restart *work_restart;
extern bool opt_trust_pool;
extern uint16_t opt_vote;
@@ -571,10 +549,6 @@ extern void diff_to_target(uint32_t *target, double diff);
extern void get_currentalgo(char* buf, int sz);
extern uint32_t device_intensity(int thr_id, const char *func, uint32_t defcount);
// bignum
double bn_convert_nbits(const uint32_t nbits);
void bn_nbits_to_uchar(const uint32_t nBits, uchar *target);
struct stratum_job {
char *job_id;
unsigned char prevhash[32];
@@ -618,13 +592,6 @@ struct stratum_ctx {
int srvtime_diff;
};
#define POK_MAX_TXS 4
#define POK_MAX_TX_SZ 16384U
struct tx {
uint8_t data[POK_MAX_TX_SZ];
uint32_t len;
};
struct work {
uint32_t data[32];
uint32_t target[8];
@@ -645,16 +612,8 @@ struct work {
uint32_t scanned_from;
uint32_t scanned_to;
/* pok getwork txs */
uint32_t tx_count;
struct tx txs[POK_MAX_TXS];
};
#define POK_BOOL_MASK 0x00008000
#define POK_DATA_MASK 0xFFFF0000
#define MAX_POOLS 8
struct pool_infos {
uint8_t id;
@@ -662,7 +621,6 @@ struct pool_infos {
#define POOL_GETWORK 1
#define POOL_STRATUM 2
#define POOL_LONGPOLL 4
#define POOL_ETHER 128
uint8_t type;
#define POOL_ST_DEFINED 1
#define POOL_ST_VALID 2
@@ -700,24 +658,18 @@ extern struct pool_infos pools[MAX_POOLS];
extern int num_pools;
extern volatile int cur_pooln;
void pool_init_defaults(void);
int pool_get_first_valid(int startfrom);
void pool_set_creds(int pooln);
void pool_set_attr(int pooln, const char* key, char* arg);
bool pool_switch_url(char *params);
bool pool_switch(int pooln);
bool pool_switch_next(void);
int pool_get_first_valid(int startfrom);
bool parse_pool_array(json_t *obj);
void pool_dump_infos(void);
json_t * json_rpc_call_pool(CURL *curl, struct pool_infos*,
const char *req, bool lp_scan, bool lp, int *err);
json_t * json_rpc_longpoll(CURL *curl, char *lp_url, struct pool_infos*,
const char *req, int *err);
bool ether_getwork(CURL *curl, struct pool_infos *pool, struct work *work);
bool ether_submitwork(CURL *curl, struct pool_infos *pool, struct work *work);
bool stratum_socket_full(struct stratum_ctx *sctx, int timeout);
bool stratum_send_line(struct stratum_ctx *sctx, char *s);
char *stratum_recv_line(struct stratum_ctx *sctx);
@@ -779,9 +731,8 @@ void applog_compare_hash(unsigned char *hash, unsigned char *hash2);
void print_hash_tests(void);
void animehash(void *state, const void *input);
void blake256hash(void *output, const void *input, int8_t rounds);
void c11hash(void *output, const void *input);
void deephash(void *state, const void *input);
void luffa_hash(void *state, const void *input);
void doomhash(void *state, const void *input);
void fresh_hash(void *state, const void *input);
void fugue256_hash(unsigned char* output, const unsigned char* input, int len);
void heavycoin_hash(unsigned char* output, const unsigned char* input, int len);
@@ -793,6 +744,7 @@ void myriadhash(void *state, const void *input);
void neoscrypt(uchar *output, const uchar *input, uint32_t profile);
void nist5hash(void *state, const void *input);
void pentablakehash(void *output, const void *input);
void pluckhash(uint32_t *hash, const uint32_t *data, uchar *hashbuffer, const int N);
void quarkhash(void *state, const void *input);
void qubithash(void *state, const void *input);
void scrypthash(void* output, const void* input);
@@ -800,7 +752,6 @@ void scryptjane_hash(void* output, const void* input);
void skeincoinhash(void *output, const void *input);
void skein2hash(void *output, const void *input);
void s3hash(void *output, const void *input);
void whirlxHash(void *state, const void *input);
void x11hash(void *output, const void *input);
void x13hash(void *output, const void *input);
void x14hash(void *output, const void *input);
+1
View File
@@ -3,6 +3,7 @@
#include <cuda_runtime.h>
#include <openssl/sha.h>
#include "uint256.h"
#include "sph/sph_groestl.h"
#include "miner.h"
+63 -448
View File
@@ -15,11 +15,9 @@
*
*/
#include <errno.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include "miner.h"
#include "nvml.h"
@@ -35,15 +33,6 @@ extern char driver_version[32];
static uint32_t device_bus_ids[MAX_GPUS] = { 0 };
extern uint32_t device_gpu_clocks[MAX_GPUS];
extern uint32_t device_mem_clocks[MAX_GPUS];
extern uint32_t device_plimit[MAX_GPUS];
extern int8_t device_pstate[MAX_GPUS];
uint32_t clock_prev[MAX_GPUS] = { 0 };
uint32_t clock_prev_mem[MAX_GPUS] = { 0 };
uint32_t limit_prev[MAX_GPUS] = { 0 };
/*
* Wrappers to emulate dlopen() on other systems like Windows
*/
@@ -118,12 +107,14 @@ nvml_handle * nvml_create()
nvmlh->nvml_dll = nvml_dll;
nvmlh->nvmlInit = (nvmlReturn_t (*)(void)) wrap_dlsym(nvmlh->nvml_dll, "nvmlInit_v2");
if (!nvmlh->nvmlInit)
nvmlh->nvmlInit = (nvmlReturn_t (*)(void)) wrap_dlsym(nvmlh->nvml_dll, "nvmlInit");
nvmlh->nvmlDeviceGetCount = (nvmlReturn_t (*)(int *)) wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetCount_v2");
if (!nvmlh->nvmlDeviceGetCount)
nvmlh->nvmlDeviceGetCount = (nvmlReturn_t (*)(int *)) wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetCount");
nvmlh->nvmlInit = (nvmlReturn_t (*)(void))
wrap_dlsym(nvmlh->nvml_dll, "nvmlInit_v2");
if (!nvmlh->nvmlInit) {
nvmlh->nvmlInit = (nvmlReturn_t (*)(void))
wrap_dlsym(nvmlh->nvml_dll, "nvmlInit");
}
nvmlh->nvmlDeviceGetCount = (nvmlReturn_t (*)(int *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetCount_v2");
nvmlh->nvmlDeviceGetHandleByIndex = (nvmlReturn_t (*)(int, nvmlDevice_t *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetHandleByIndex_v2");
nvmlh->nvmlDeviceGetAPIRestriction = (nvmlReturn_t (*)(nvmlDevice_t, nvmlRestrictedAPI_t, nvmlEnableState_t *))
@@ -136,37 +127,10 @@ nvml_handle * nvml_create()
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetApplicationsClock");
nvmlh->nvmlDeviceSetApplicationsClocks = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int mem, unsigned int gpu))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceSetApplicationsClocks");
nvmlh->nvmlDeviceResetApplicationsClocks = (nvmlReturn_t (*)(nvmlDevice_t))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceResetApplicationsClocks");
nvmlh->nvmlDeviceGetSupportedGraphicsClocks = (nvmlReturn_t (*)(nvmlDevice_t, uint32_t mem, uint32_t *num, uint32_t *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetSupportedGraphicsClocks");
nvmlh->nvmlDeviceGetSupportedMemoryClocks = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *count, unsigned int *clocksMHz))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetSupportedMemoryClocks");
nvmlh->nvmlDeviceGetClockInfo = (nvmlReturn_t (*)(nvmlDevice_t, nvmlClockType_t, unsigned int *clock))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetClockInfo");
nvmlh->nvmlDeviceGetMaxClockInfo = (nvmlReturn_t (*)(nvmlDevice_t, nvmlClockType_t, unsigned int *clock))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetMaxClockInfo");
nvmlh->nvmlDeviceGetPciInfo = (nvmlReturn_t (*)(nvmlDevice_t, nvmlPciInfo_t *)) wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPciInfo_v2");
if (!nvmlh->nvmlDeviceGetPciInfo)
nvmlh->nvmlDeviceGetPciInfo = (nvmlReturn_t (*)(nvmlDevice_t, nvmlPciInfo_t *)) wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPciInfo");
nvmlh->nvmlDeviceGetCurrPcieLinkGeneration = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *gen))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetCurrPcieLinkGeneration");
nvmlh->nvmlDeviceGetCurrPcieLinkWidth = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *width))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetCurrPcieLinkWidth");
nvmlh->nvmlDeviceGetMaxPcieLinkGeneration = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *gen))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetMaxPcieLinkGeneration");
nvmlh->nvmlDeviceGetMaxPcieLinkWidth = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *width))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetMaxPcieLinkWidth");
nvmlh->nvmlDeviceGetPowerUsage = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPowerUsage");
nvmlh->nvmlDeviceGetPowerManagementDefaultLimit = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *limit))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPowerManagementDefaultLimit");
nvmlh->nvmlDeviceGetPowerManagementLimit = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *limit))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPowerManagementLimit");
nvmlh->nvmlDeviceGetPowerManagementLimitConstraints = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *min, unsigned int *max))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPowerManagementLimitConstraints");
nvmlh->nvmlDeviceSetPowerManagementLimit = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int limit))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceSetPowerManagementLimit");
nvmlh->nvmlDeviceGetPciInfo = (nvmlReturn_t (*)(nvmlDevice_t, nvmlPciInfo_t *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPciInfo");
nvmlh->nvmlDeviceGetName = (nvmlReturn_t (*)(nvmlDevice_t, char *, int))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetName");
nvmlh->nvmlDeviceGetTemperature = (nvmlReturn_t (*)(nvmlDevice_t, int, unsigned int *))
@@ -174,7 +138,7 @@ nvml_handle * nvml_create()
nvmlh->nvmlDeviceGetFanSpeed = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetFanSpeed");
nvmlh->nvmlDeviceGetPerformanceState = (nvmlReturn_t (*)(nvmlDevice_t, int *))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPerformanceState"); /* or nvmlDeviceGetPowerState */
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPowerUsage");
nvmlh->nvmlDeviceGetSerial = (nvmlReturn_t (*)(nvmlDevice_t, char *, unsigned int))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetSerial");
nvmlh->nvmlDeviceGetUUID = (nvmlReturn_t (*)(nvmlDevice_t, char *, unsigned int))
@@ -187,26 +151,17 @@ nvml_handle * nvml_create()
wrap_dlsym(nvmlh->nvml_dll, "nvmlErrorString");
nvmlh->nvmlShutdown = (nvmlReturn_t (*)())
wrap_dlsym(nvmlh->nvml_dll, "nvmlShutdown");
// v331
nvmlh->nvmlDeviceGetEnforcedPowerLimit = (nvmlReturn_t (*)(nvmlDevice_t, unsigned int *limit))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetEnforcedPowerLimit");
// v340
/* NVML_ERROR_NOT_SUPPORTED
nvmlh->nvmlDeviceGetAutoBoostedClocksEnabled = (nvmlReturn_t (*)(nvmlDevice_t, nvmlEnableState_t *isEnabled, nvmlEnableState_t *defaultIsEnabled))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetAutoBoostedClocksEnabled");
nvmlh->nvmlDeviceSetAutoBoostedClocksEnabled = (nvmlReturn_t (*)(nvmlDevice_t, nvmlEnableState_t enabled))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceSetAutoBoostedClocksEnabled"); */
// v346
nvmlh->nvmlDeviceGetPcieThroughput = (nvmlReturn_t (*)(nvmlDevice_t, nvmlPcieUtilCounter_t, unsigned int *value))
wrap_dlsym(nvmlh->nvml_dll, "nvmlDeviceGetPcieThroughput");
if (nvmlh->nvmlInit == NULL ||
nvmlh->nvmlShutdown == NULL ||
nvmlh->nvmlErrorString == NULL ||
nvmlh->nvmlSystemGetDriverVersion == NULL ||
nvmlh->nvmlDeviceGetCount == NULL ||
nvmlh->nvmlDeviceGetHandleByIndex == NULL ||
nvmlh->nvmlDeviceGetPciInfo == NULL ||
nvmlh->nvmlDeviceGetName == NULL)
nvmlh->nvmlDeviceGetName == NULL ||
nvmlh->nvmlDeviceGetTemperature == NULL ||
nvmlh->nvmlDeviceGetFanSpeed == NULL)
{
if (opt_debug)
applog(LOG_DEBUG, "Failed to obtain required NVML function pointers");
@@ -216,8 +171,7 @@ nvml_handle * nvml_create()
}
nvmlh->nvmlInit();
if (nvmlh->nvmlSystemGetDriverVersion)
nvmlh->nvmlSystemGetDriverVersion(driver_version, sizeof(driver_version));
nvmlh->nvmlSystemGetDriverVersion(driver_version, sizeof(driver_version));
nvmlh->nvmlDeviceGetCount(&nvmlh->nvml_gpucount);
/* Query CUDA device count, in case it doesn't agree with NVML, since */
@@ -253,7 +207,7 @@ nvml_handle * nvml_create()
nvmlh->nvml_pci_domain_id[i] = pciinfo.domain;
nvmlh->nvml_pci_bus_id[i] = pciinfo.bus;
nvmlh->nvml_pci_device_id[i] = pciinfo.device;
nvmlh->nvml_pci_subsys_id[i] = pciinfo.pci_subsystem_id;
nvmlh->nvml_pci_subsys_id[i] = pciinfo.pci_device_id;
nvmlh->app_clocks[i] = NVML_FEATURE_UNKNOWN;
if (nvmlh->nvmlDeviceSetAPIRestriction) {
@@ -264,6 +218,21 @@ nvml_handle * nvml_create()
if (nvmlh->nvmlDeviceGetAPIRestriction) {
nvmlh->nvmlDeviceGetAPIRestriction(nvmlh->devs[i], NVML_RESTRICTED_API_SET_APPLICATION_CLOCKS,
&nvmlh->app_clocks[i]);
if (nvmlh->app_clocks[i] == NVML_FEATURE_ENABLED && opt_debug) {
applog(LOG_DEBUG, "NVML application clock feature is allowed");
#if 0
uint32_t mem;
nvmlReturn_t rc;
rc = nvmlh->nvmlDeviceGetDefaultApplicationsClock(nvmlh->devs[i], NVML_CLOCK_MEM, &mem);
if (rc == NVML_SUCCESS)
applog(LOG_DEBUG, "nvmlDeviceGetDefaultApplicationsClock: mem %u", mem);
else
applog(LOG_DEBUG, "nvmlDeviceGetDefaultApplicationsClock: %s", nvmlh->nvmlErrorString(rc));
rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[i], mem, 1228000);
if (rc != NVML_SUCCESS)
applog(LOG_DEBUG, "nvmlDeviceSetApplicationsClocks: %s", nvmlh->nvmlErrorString(rc));
#endif
}
}
}
@@ -282,7 +251,7 @@ nvml_handle * nvml_create()
(nvmlh->nvml_pci_bus_id[j] == (uint32_t) props.pciBusID) &&
(nvmlh->nvml_pci_device_id[j] == (uint32_t) props.pciDeviceID)) {
if (opt_debug)
applog(LOG_DEBUG, "CUDA GPU %d matches NVML GPU %d by busId %u",
applog(LOG_DEBUG, "CUDA GPU#%d matches NVML GPU %d by busId %u",
i, j, (uint32_t) props.pciBusID);
nvmlh->nvml_cuda_device_id[j] = i;
nvmlh->cuda_nvml_device_id[i] = j;
@@ -294,234 +263,6 @@ nvml_handle * nvml_create()
return nvmlh;
}
/* apply config clocks to an used device */
int nvml_set_clocks(nvml_handle *nvmlh, int dev_id)
{
nvmlReturn_t rc;
uint32_t gpu_clk = 0, mem_clk = 0;
int n = nvmlh->cuda_nvml_device_id[dev_id];
if (n < 0 || n >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!device_gpu_clocks[dev_id] && !device_mem_clocks[dev_id])
return 0; // nothing to do
if (nvmlh->app_clocks[n] != NVML_FEATURE_ENABLED) {
applog(LOG_WARNING, "GPU #%d: NVML application clock feature is not allowed!", dev_id);
return -EPERM;
}
uint32_t mem_prev = clock_prev_mem[dev_id];
if (!mem_prev)
nvmlh->nvmlDeviceGetApplicationsClock(nvmlh->devs[n], NVML_CLOCK_MEM, &mem_prev);
uint32_t gpu_prev = clock_prev[dev_id];
if (!gpu_prev)
nvmlh->nvmlDeviceGetApplicationsClock(nvmlh->devs[n], NVML_CLOCK_GRAPHICS, &gpu_prev);
nvmlh->nvmlDeviceGetDefaultApplicationsClock(nvmlh->devs[n], NVML_CLOCK_MEM, &mem_clk);
rc = nvmlh->nvmlDeviceGetDefaultApplicationsClock(nvmlh->devs[n], NVML_CLOCK_GRAPHICS, &gpu_clk);
if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: unable to query application clocks", dev_id);
return -EINVAL;
}
if (opt_debug)
applog(LOG_DEBUG, "GPU #%d: default application clocks are %u/%u", dev_id, mem_clk, gpu_clk);
// get application config values
if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id];
if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id];
// these functions works for the 960 and the 970 (346.72+), not for the 750 Ti
uint32_t nclocks = 0, clocks[127] = { 0 };
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL);
nclocks = min(nclocks, 127);
if (nclocks)
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, clocks);
for (int8_t u=0; u < nclocks; u++) {
// ordered by pstate (so highest is first memory clock - P0)
if (clocks[u] <= mem_clk) {
mem_clk = clocks[u];
break;
}
}
nclocks = 0;
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL);
nclocks = min(nclocks, 127);
if (nclocks)
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, clocks);
for (uint8_t u=0; u < nclocks; u++) {
// ordered desc, so get first
if (clocks[u] <= gpu_clk) {
gpu_clk = clocks[u];
break;
}
}
rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk);
if (rc == NVML_SUCCESS)
applog(LOG_INFO, "GPU #%d: application clocks set to %u/%u", dev_id, mem_clk, gpu_clk);
else {
applog(LOG_WARNING, "GPU #%d: %u/%u - %s", dev_id, mem_clk, gpu_clk, nvmlh->nvmlErrorString(rc));
return -1;
}
// store previous clocks for reset on exit (or during wait...)
clock_prev[dev_id] = gpu_prev;
clock_prev_mem[dev_id] = mem_prev;
return 1;
}
/* reset default app clocks and limits on exit */
int nvml_reset_clocks(nvml_handle *nvmlh, int dev_id)
{
int ret = 0;
nvmlReturn_t rc;
uint32_t gpu_clk = 0, mem_clk = 0;
int n = nvmlh->cuda_nvml_device_id[dev_id];
if (n < 0 || n >= nvmlh->nvml_gpucount)
return -ENODEV;
if (clock_prev[dev_id]) {
rc = nvmlh->nvmlDeviceResetApplicationsClocks(nvmlh->devs[n]);
if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: unable to reset application clocks", dev_id);
}
clock_prev[dev_id] = 0;
ret = 1;
}
if (limit_prev[dev_id]) {
uint32_t plimit = limit_prev[dev_id];
if (nvmlh->nvmlDeviceGetPowerManagementDefaultLimit && !plimit) {
rc = nvmlh->nvmlDeviceGetPowerManagementDefaultLimit(nvmlh->devs[n], &plimit);
} else if (plimit) {
rc = NVML_SUCCESS;
}
if (rc == NVML_SUCCESS)
nvmlh->nvmlDeviceSetPowerManagementLimit(nvmlh->devs[n], plimit);
ret = 1;
}
return ret;
}
/**
* Set power state of a device (9xx)
* Code is similar as clocks one, which allow the change of the pstate
*/
int nvml_set_pstate(nvml_handle *nvmlh, int dev_id)
{
nvmlReturn_t rc;
uint32_t gpu_clk = 0, mem_clk = 0;
int n = nvmlh->cuda_nvml_device_id[dev_id];
if (n < 0 || n >= nvmlh->nvml_gpucount)
return -ENODEV;
if (device_pstate[dev_id] < 0)
return 0;
if (nvmlh->app_clocks[n] != NVML_FEATURE_ENABLED) {
applog(LOG_WARNING, "GPU #%d: NVML app. clock feature is not allowed!", dev_id);
return -EPERM;
}
nvmlh->nvmlDeviceGetDefaultApplicationsClock(nvmlh->devs[n], NVML_CLOCK_MEM, &mem_clk);
rc = nvmlh->nvmlDeviceGetDefaultApplicationsClock(nvmlh->devs[n], NVML_CLOCK_GRAPHICS, &gpu_clk);
if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: unable to query application clocks", dev_id);
return -EINVAL;
}
// get application config values
if (device_mem_clocks[dev_id]) mem_clk = device_mem_clocks[dev_id];
if (device_gpu_clocks[dev_id]) gpu_clk = device_gpu_clocks[dev_id];
// these functions works for the 960 and the 970 (346.72+), not for the 750 Ti
uint32_t nclocks = 0, clocks[127] = { 0 };
int8_t wanted_pstate = device_pstate[dev_id];
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, NULL);
nclocks = min(nclocks, 127);
if (nclocks)
nvmlh->nvmlDeviceGetSupportedMemoryClocks(nvmlh->devs[n], &nclocks, clocks);
for (uint8_t u=0; u < nclocks; u++) {
// ordered by pstate (so highest P0 first)
if (u == wanted_pstate) {
mem_clk = clocks[u];
break;
}
}
nclocks = 0;
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, NULL);
nclocks = min(nclocks, 127);
if (nclocks)
nvmlh->nvmlDeviceGetSupportedGraphicsClocks(nvmlh->devs[n], mem_clk, &nclocks, clocks);
for (uint8_t u=0; u < nclocks; u++) {
// ordered desc, so get first
if (clocks[u] <= gpu_clk) {
gpu_clk = clocks[u];
break;
}
}
rc = nvmlh->nvmlDeviceSetApplicationsClocks(nvmlh->devs[n], mem_clk, gpu_clk);
if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: pstate %s", dev_id, nvmlh->nvmlErrorString(rc));
return -1;
}
if (!opt_quiet)
applog(LOG_INFO, "GPU #%d: app clocks set to P%d (%u/%u)", dev_id, (int) wanted_pstate, mem_clk, gpu_clk);
clock_prev[dev_id] = 1;
return 1;
}
int nvml_set_plimit(nvml_handle *nvmlh, int dev_id)
{
nvmlReturn_t rc = NVML_ERROR_UNKNOWN;
uint32_t gpu_clk = 0, mem_clk = 0;
int n = nvmlh->cuda_nvml_device_id[dev_id];
if (n < 0 || n >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!device_plimit[dev_id])
return 0; // nothing to do
if (!nvmlh->nvmlDeviceSetPowerManagementLimit)
return -ENOSYS;
uint32_t plimit = device_plimit[dev_id] * 1000;
uint32_t pmin = 1000, pmax = 0, prev_limit = 0;
if (nvmlh->nvmlDeviceGetPowerManagementLimitConstraints)
rc = nvmlh->nvmlDeviceGetPowerManagementLimitConstraints(nvmlh->devs[n], &pmin, &pmax);
if (rc != NVML_SUCCESS) {
if (!nvmlh->nvmlDeviceGetPowerManagementLimit)
return -ENOSYS;
}
nvmlh->nvmlDeviceGetPowerManagementLimit(nvmlh->devs[n], &prev_limit);
if (!pmax) pmax = prev_limit;
plimit = min(plimit, pmax);
plimit = max(plimit, pmin);
rc = nvmlh->nvmlDeviceSetPowerManagementLimit(nvmlh->devs[n], plimit);
if (rc != NVML_SUCCESS) {
applog(LOG_WARNING, "GPU #%d: plimit %s", dev_id, nvmlh->nvmlErrorString(rc));
return -1;
}
if (!opt_quiet) {
applog(LOG_INFO, "GPU #%d: power limit set to %uW (allowed range is %u-%u)",
dev_id, plimit/1000U, pmin/1000U, pmax/1000U);
}
limit_prev[dev_id] = prev_limit;
return 1;
}
int nvml_get_gpucount(nvml_handle *nvmlh, int *gpucount)
{
*gpucount = nvmlh->nvml_gpucount;
@@ -539,10 +280,7 @@ int nvml_get_gpu_name(nvml_handle *nvmlh, int cudaindex, char *namebuf, int bufs
{
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetName)
return -ENOSYS;
return -1;
if (nvmlh->nvmlDeviceGetName(nvmlh->devs[gpuindex], namebuf, bufsize) != NVML_SUCCESS)
return -1;
@@ -556,10 +294,7 @@ int nvml_get_tempC(nvml_handle *nvmlh, int cudaindex, unsigned int *tempC)
nvmlReturn_t rc;
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetTemperature)
return -ENOSYS;
return -1;
rc = nvmlh->nvmlDeviceGetTemperature(nvmlh->devs[gpuindex], 0u /* NVML_TEMPERATURE_GPU */, tempC);
if (rc != NVML_SUCCESS) {
@@ -575,10 +310,7 @@ int nvml_get_fanpcnt(nvml_handle *nvmlh, int cudaindex, unsigned int *fanpcnt)
nvmlReturn_t rc;
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetFanSpeed)
return -ENOSYS;
return -1;
rc = nvmlh->nvmlDeviceGetFanSpeed(nvmlh->devs[gpuindex], fanpcnt);
if (rc != NVML_SUCCESS) {
@@ -593,15 +325,12 @@ int nvml_get_power_usage(nvml_handle *nvmlh, int cudaindex, unsigned int *milliw
{
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetPowerUsage)
return -ENOSYS;
return -1;
nvmlReturn_t res = nvmlh->nvmlDeviceGetPowerUsage(nvmlh->devs[gpuindex], milliwatts);
if (res != NVML_SUCCESS) {
//if (opt_debug)
// applog(LOG_DEBUG, "nvmlDeviceGetPowerUsage: %s", nvmlh->nvmlErrorString(res));
if (opt_debug)
applog(LOG_DEBUG, "nvmlDeviceGetPowerUsage: %s", nvmlh->nvmlErrorString(res));
return -1;
}
@@ -613,10 +342,7 @@ int nvml_get_pstate(nvml_handle *nvmlh, int cudaindex, int *pstate)
{
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetPerformanceState)
return -ENOSYS;
return -1;
nvmlReturn_t res = nvmlh->nvmlDeviceGetPerformanceState(nvmlh->devs[gpuindex], pstate);
if (res != NVML_SUCCESS) {
@@ -632,7 +358,7 @@ int nvml_get_busid(nvml_handle *nvmlh, int cudaindex, int *busid)
{
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
return -1;
(*busid) = nvmlh->nvml_pci_bus_id[gpuindex];
return 0;
@@ -645,17 +371,13 @@ int nvml_get_serial(nvml_handle *nvmlh, int cudaindex, char *sn, int maxlen)
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
nvmlReturn_t res;
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
return -1;
if (nvmlh->nvmlDeviceGetSerial) {
res = nvmlh->nvmlDeviceGetSerial(nvmlh->devs[gpuindex], sn, maxlen);
if (res == NVML_SUCCESS)
return 0;
res = nvmlh->nvmlDeviceGetSerial(nvmlh->devs[gpuindex], sn, maxlen);
if (res == NVML_SUCCESS) {
return 0;
}
if (!nvmlh->nvmlDeviceGetUUID)
return -ENOSYS;
// nvmlDeviceGetUUID: GPU-f2bd642c-369f-5a14-e0b4-0d22dfe9a1fc
// use a part of uuid to generate an unique serial
// todo: check if there is vendor id is inside
@@ -676,10 +398,7 @@ int nvml_get_bios(nvml_handle *nvmlh, int cudaindex, char *desc, int maxlen)
uint32_t subids = 0;
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
if (!nvmlh->nvmlDeviceGetVbiosVersion)
return -ENOSYS;
return -1;
nvmlReturn_t res = nvmlh->nvmlDeviceGetVbiosVersion(nvmlh->devs[gpuindex], desc, maxlen);
if (res != NVML_SUCCESS) {
@@ -690,16 +409,16 @@ int nvml_get_bios(nvml_handle *nvmlh, int cudaindex, char *desc, int maxlen)
return 0;
}
int nvml_get_info(nvml_handle *nvmlh, int cudaindex, uint16_t &vid, uint16_t &pid)
int nvml_get_info(nvml_handle *nvmlh, int cudaindex, uint16_t *vid, uint16_t *pid)
{
uint32_t subids = 0;
int gpuindex = nvmlh->cuda_nvml_device_id[cudaindex];
if (gpuindex < 0 || gpuindex >= nvmlh->nvml_gpucount)
return -ENODEV;
return -1;
subids = nvmlh->nvml_pci_subsys_id[gpuindex];
pid = subids >> 16;
vid = subids & 0xFFFF;
(*pid) = subids >> 16;
(*vid) = subids & 0xFFFF;
return 0;
}
@@ -739,7 +458,7 @@ int nvapi_temperature(unsigned int devNum, unsigned int *temperature)
NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
NV_GPU_THERMAL_SETTINGS thermal;
thermal.version = NV_GPU_THERMAL_SETTINGS_VER;
@@ -762,7 +481,7 @@ int nvapi_fanspeed(unsigned int devNum, unsigned int *speed)
NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
NvU32 fanspeed = 0;
ret = NvAPI_GPU_GetTachReading(phys[devNum], &fanspeed);
@@ -784,7 +503,7 @@ int nvapi_getpstate(unsigned int devNum, unsigned int *power)
NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
NV_GPU_PERF_PSTATE_ID CurrentPstate = NVAPI_GPU_PERF_PSTATE_UNDEFINED; /* 16 */
ret = NvAPI_GPU_GetCurrentPstate(phys[devNum], &CurrentPstate);
@@ -809,7 +528,7 @@ int nvapi_getusage(unsigned int devNum, unsigned int *pct)
NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
NV_GPU_DYNAMIC_PSTATES_INFO_EX info;
info.version = NV_GPU_DYNAMIC_PSTATES_INFO_EX_VER;
@@ -829,13 +548,13 @@ int nvapi_getusage(unsigned int devNum, unsigned int *pct)
return 0;
}
int nvapi_getinfo(unsigned int devNum, uint16_t &vid, uint16_t &pid)
int nvapi_getinfo(unsigned int devNum, uint16_t *vid, uint16_t *pid)
{
NvAPI_Status ret;
NvU32 pDeviceId, pSubSystemId, pRevisionId, pExtDeviceId;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
ret = NvAPI_GPU_GetPCIIdentifiers(phys[devNum], &pDeviceId, &pSubSystemId, &pRevisionId, &pExtDeviceId);
if (ret != NVAPI_OK) {
@@ -846,8 +565,8 @@ int nvapi_getinfo(unsigned int devNum, uint16_t &vid, uint16_t &pid)
return -1;
}
pid = pDeviceId >> 16;
vid = pDeviceId & 0xFFFF;
(*pid) = pDeviceId >> 16;
(*vid) = pDeviceId & 0xFFFF;
return 0;
}
@@ -856,7 +575,7 @@ int nvapi_getserial(unsigned int devNum, char *serial, unsigned int maxlen)
{
// NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
sprintf(serial, "");
@@ -880,7 +599,7 @@ int nvapi_getbios(unsigned int devNum, char *desc, unsigned int maxlen)
{
NvAPI_Status ret;
if (devNum >= nvapi_dev_cnt)
return -ENODEV;
return -1;
if (maxlen < 64) // Short String
return -1;
@@ -957,7 +676,7 @@ int nvapi_init()
NvAPI_ShortString str;
ret = NvAPI_SYS_GetDriverAndBranchVersion(&udv, str);
if (ret == NVAPI_OK) {
sprintf(driver_version,"%d.%02d", udv / 100, udv % 100);
sprintf(driver_version,"%d.%d", udv/100, udv % 100);
}
return 0;
@@ -1066,107 +785,9 @@ unsigned int gpu_power(struct cgpu_info *gpu)
return mw;
}
static int translate_vendor_id(uint16_t vid, char *vendorname)
{
struct VENDORS {
const uint16_t vid;
const char *name;
} vendors[] = {
{ 0x1043, "ASUS" },
{ 0x107D, "Leadtek" },
{ 0x10B0, "Gainward" },
// { 0x10DE, "NVIDIA" },
{ 0x1458, "Gigabyte" },
{ 0x1462, "MSI" },
{ 0x154B, "PNY" },
{ 0x1682, "XFX" },
{ 0x196D, "Club3D" },
{ 0x19DA, "Zotac" },
{ 0x19F1, "BFG" },
{ 0x1ACC, "PoV" },
{ 0x1B4C, "KFA2" },
{ 0x3842, "EVGA" },
{ 0x7377, "Colorful" },
{ 0, "" }
};
if (!vendorname)
return -EINVAL;
for(int v=0; v < ARRAY_SIZE(vendors); v++) {
if (vid == vendors[v].vid) {
strcpy(vendorname, vendors[v].name);
return vid;
}
}
if (opt_debug)
applog(LOG_DEBUG, "nvml: Unknown vendor %04x\n", vid);
return 0;
}
#ifdef HAVE_PCIDEV
extern "C" {
#include <pci/pci.h>
}
static int linux_gpu_vendor(uint8_t pci_bus_id, char* vendorname, uint16_t &pid)
{
uint16_t subvendor = 0;
struct pci_access *pci;
struct pci_dev *dev;
uint16_t subdevice;
if (!vendorname)
return -EINVAL;
pci = pci_alloc();
if (!pci)
return -ENODEV;
pci_init(pci);
pci_scan_bus(pci);
for(dev = pci->devices; dev; dev = dev->next)
{
if (dev->bus == pci_bus_id && dev->vendor_id == 0x10DE)
{
if (!(dev->known_fields & PCI_FILL_CLASS))
pci_fill_info(dev, PCI_FILL_CLASS);
if (dev->device_class != PCI_CLASS_DISPLAY_VGA)
continue;
subvendor = pci_read_word(dev, PCI_SUBSYSTEM_VENDOR_ID);
subdevice = pci_read_word(dev, PCI_SUBSYSTEM_ID); // model
translate_vendor_id(subvendor, vendorname);
}
}
pci_cleanup(pci);
return (int) subvendor;
}
#endif
int gpu_vendor(uint8_t pci_bus_id, char *vendorname)
{
#ifdef HAVE_PCIDEV
uint16_t pid = 0;
return linux_gpu_vendor(pci_bus_id, vendorname, pid);
#else
uint16_t vid = 0, pid = 0;
if (hnvml) { // may not be initialized on start...
for (int id=0; id < hnvml->nvml_gpucount; id++) {
if (hnvml->nvml_pci_bus_id[id] == pci_bus_id) {
nvml_get_info(hnvml, id, vid, pid);
}
}
}
return translate_vendor_id(vid, vendorname);
#endif
}
int gpu_info(struct cgpu_info *gpu)
{
char vendorname[32] = { 0 };
int id = gpu->gpu_id;
uint8_t bus_id = 0;
gpu->nvml_id = -1;
gpu->nvapi_id = -1;
@@ -1176,19 +797,13 @@ int gpu_info(struct cgpu_info *gpu)
if (hnvml) {
gpu->nvml_id = (int8_t) hnvml->cuda_nvml_device_id[id];
#ifdef HAVE_PCIDEV
gpu->gpu_vid = linux_gpu_vendor(hnvml->nvml_pci_bus_id[id], vendorname, gpu->gpu_pid);
if (!gpu->gpu_vid || !gpu->gpu_pid)
nvml_get_info(hnvml, id, gpu->gpu_vid, gpu->gpu_pid);
#else
nvml_get_info(hnvml, id, gpu->gpu_vid, gpu->gpu_pid);
#endif
nvml_get_info(hnvml, id, &gpu->gpu_vid, &gpu->gpu_pid);
nvml_get_serial(hnvml, id, gpu->gpu_sn, sizeof(gpu->gpu_sn));
nvml_get_bios(hnvml, id, gpu->gpu_desc, sizeof(gpu->gpu_desc));
}
#ifdef WIN32
gpu->nvapi_id = (int8_t) nvapi_dev_map[id];
nvapi_getinfo(nvapi_dev_map[id], gpu->gpu_vid, gpu->gpu_pid);
nvapi_getinfo(nvapi_dev_map[id], &gpu->gpu_vid, &gpu->gpu_pid);
nvapi_getserial(nvapi_dev_map[id], gpu->gpu_sn, sizeof(gpu->gpu_sn));
nvapi_getbios(nvapi_dev_map[id], gpu->gpu_desc, sizeof(gpu->gpu_desc));
#endif
+37 -41
View File
@@ -66,20 +66,6 @@ enum nvmlClockType_t {
NVML_CLOCK_MEM = 2
};
enum nvmlPcieUtilCounter_t {
NVML_PCIE_UTIL_TX_BYTES = 0,
NVML_PCIE_UTIL_RX_BYTES = 1,
NVML_PCIE_UTIL_COUNT
};
enum nvmlValueType_t {
NVML_VALUE_TYPE_DOUBLE = 0,
NVML_VALUE_TYPE_UNSIGNED_INT = 1,
NVML_VALUE_TYPE_UNSIGNED_LONG = 2,
NVML_VALUE_TYPE_UNSIGNED_LONG_LONG = 3,
NVML_VALUE_TYPE_COUNT
};
#define NVML_DEVICE_SERIAL_BUFFER_SIZE 30
#define NVML_DEVICE_UUID_BUFFER_SIZE 80
#define NVML_DEVICE_VBIOS_VERSION_BUFFER_SIZE 32
@@ -108,20 +94,8 @@ typedef struct {
nvmlReturn_t (*nvmlDeviceGetDefaultApplicationsClock)(nvmlDevice_t, nvmlClockType_t, unsigned int *);
nvmlReturn_t (*nvmlDeviceGetApplicationsClock)(nvmlDevice_t, nvmlClockType_t, unsigned int *);
nvmlReturn_t (*nvmlDeviceSetApplicationsClocks)(nvmlDevice_t, unsigned int, unsigned int);
nvmlReturn_t (*nvmlDeviceResetApplicationsClocks)(nvmlDevice_t);
nvmlReturn_t (*nvmlDeviceGetSupportedGraphicsClocks)(nvmlDevice_t, uint32_t mem, uint32_t *num, uint32_t *arr);
nvmlReturn_t (*nvmlDeviceGetSupportedMemoryClocks)(nvmlDevice_t, unsigned int *count, unsigned int *clocksMHz);
nvmlReturn_t (*nvmlDeviceGetClockInfo)(nvmlDevice_t, nvmlClockType_t, unsigned int *);
nvmlReturn_t (*nvmlDeviceGetMaxClockInfo)(nvmlDevice_t, nvmlClockType_t, unsigned int *);
nvmlReturn_t (*nvmlDeviceGetPowerManagementDefaultLimit)(nvmlDevice_t, unsigned int *limit);
nvmlReturn_t (*nvmlDeviceGetPowerManagementLimit)(nvmlDevice_t, unsigned int *limit);
nvmlReturn_t (*nvmlDeviceGetPowerManagementLimitConstraints)(nvmlDevice_t, unsigned int *min, unsigned int *max);
nvmlReturn_t (*nvmlDeviceSetPowerManagementLimit)(nvmlDevice_t device, unsigned int limit);
nvmlReturn_t (*nvmlDeviceGetPciInfo)(nvmlDevice_t, nvmlPciInfo_t *);
nvmlReturn_t (*nvmlDeviceGetCurrPcieLinkGeneration)(nvmlDevice_t device, unsigned int *gen);
nvmlReturn_t (*nvmlDeviceGetCurrPcieLinkWidth)(nvmlDevice_t device, unsigned int *width);
nvmlReturn_t (*nvmlDeviceGetMaxPcieLinkGeneration)(nvmlDevice_t device, unsigned int *gen);
nvmlReturn_t (*nvmlDeviceGetMaxPcieLinkWidth)(nvmlDevice_t device, unsigned int *width);
nvmlReturn_t (*nvmlDeviceGetName)(nvmlDevice_t, char *, int);
nvmlReturn_t (*nvmlDeviceGetTemperature)(nvmlDevice_t, int, unsigned int *);
nvmlReturn_t (*nvmlDeviceGetFanSpeed)(nvmlDevice_t, unsigned int *);
@@ -133,15 +107,6 @@ typedef struct {
nvmlReturn_t (*nvmlSystemGetDriverVersion)(char *version, unsigned int len);
char* (*nvmlErrorString)(nvmlReturn_t);
nvmlReturn_t (*nvmlShutdown)(void);
// v331
nvmlReturn_t (*nvmlDeviceGetEnforcedPowerLimit)(nvmlDevice_t, unsigned int *limit);
// v340
//nvmlReturn_t (*nvmlDeviceGetCpuAffinity)(nvmlDevice_t, unsigned int cpuSetSize, unsigned long* cpuSet);
//nvmlReturn_t (*nvmlDeviceSetCpuAffinity)(nvmlDevice_t);
//nvmlReturn_t (*nvmlDeviceGetAutoBoostedClocksEnabled)(nvmlDevice_t, nvmlEnableState_t *isEnabled, nvmlEnableState_t *defaultIsEnabled);
//nvmlReturn_t (*nvmlDeviceSetAutoBoostedClocksEnabled)(nvmlDevice_t, nvmlEnableState_t enabled);
// v346
nvmlReturn_t (*nvmlDeviceGetPcieThroughput)(nvmlDevice_t, nvmlPcieUtilCounter_t, unsigned int *value);
} nvml_handle;
@@ -153,11 +118,43 @@ int nvml_destroy(nvml_handle *nvmlh);
*/
int nvml_get_gpucount(nvml_handle *nvmlh, int *gpucount);
int nvml_set_plimit(nvml_handle *nvmlh, int dev_id);
int nvml_set_pstate(nvml_handle *nvmlh, int dev_id);
/*
* Query the number of GPUs seen by CUDA
*/
int cuda_get_gpucount(nvml_handle *nvmlh, int *gpucount);
int nvml_set_clocks(nvml_handle *nvmlh, int dev_id);
int nvml_reset_clocks(nvml_handle *nvmlh, int dev_id);
/*
* query the name of the GPU model from the CUDA device ID
*
*/
int nvml_get_gpu_name(nvml_handle *nvmlh,
int gpuindex,
char *namebuf,
int bufsize);
/*
* Query the current GPU temperature (Celsius), from the CUDA device ID
*/
int nvml_get_tempC(nvml_handle *nvmlh,
int gpuindex, unsigned int *tempC);
/*
* Query the current GPU fan speed (percent) from the CUDA device ID
*/
int nvml_get_fanpcnt(nvml_handle *nvmlh,
int gpuindex, unsigned int *fanpcnt);
/*
* Query the current GPU power usage in millwatts from the CUDA device ID
*
* This feature is only available on recent GPU generations and may be
* limited in some cases only to Tesla series GPUs.
* If the query is run on an unsupported GPU, this routine will return -1.
*/
int nvml_get_power_usage(nvml_handle *nvmlh,
int gpuindex,
unsigned int *milliwatts);
/* api functions */
@@ -165,13 +162,12 @@ unsigned int gpu_fanpercent(struct cgpu_info *gpu);
unsigned int gpu_fanrpm(struct cgpu_info *gpu);
float gpu_temp(struct cgpu_info *gpu);
unsigned int gpu_power(struct cgpu_info *gpu);
unsigned int gpu_usage(struct cgpu_info *gpu);
int gpu_pstate(struct cgpu_info *gpu);
int gpu_busid(struct cgpu_info *gpu);
/* pid/vid, sn and bios rev */
int gpu_info(struct cgpu_info *gpu);
int gpu_vendor(uint8_t pci_bus_id, char *vendorname);
/* nvapi functions */
#ifdef WIN32
+573
View File
@@ -0,0 +1,573 @@
/*
* "pluck" kernel implementation.
*
* ==========================(LICENSE BEGIN)============================
*
* Copyright (c) 2015 djm34
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*
* ===========================(LICENSE END)=============================
*
* @author djm34
* @author tpruvot
*/
#include <stdio.h>
#include <stdint.h>
#include <memory.h>
#include "cuda_vector.h"
#include "miner.h"
uint32_t *d_PlNonce[MAX_GPUS];
__device__ uint8_t * hashbuffer;
__constant__ uint32_t pTarget[8];
__constant__ uint32_t c_data[20];
#define HASH_MEMORY_8bit 131072
#define HASH_MEMORY_32bit 32768
#define HASH_MEMORY 4096
static __constant__ uint32_t H256[8] = {
0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
};
static __constant__ uint32_t Ksha[64] = {
0x428A2F98, 0x71374491, 0xB5C0FBCF, 0xE9B5DBA5,
0x3956C25B, 0x59F111F1, 0x923F82A4, 0xAB1C5ED5,
0xD807AA98, 0x12835B01, 0x243185BE, 0x550C7DC3,
0x72BE5D74, 0x80DEB1FE, 0x9BDC06A7, 0xC19BF174,
0xE49B69C1, 0xEFBE4786, 0x0FC19DC6, 0x240CA1CC,
0x2DE92C6F, 0x4A7484AA, 0x5CB0A9DC, 0x76F988DA,
0x983E5152, 0xA831C66D, 0xB00327C8, 0xBF597FC7,
0xC6E00BF3, 0xD5A79147, 0x06CA6351, 0x14292967,
0x27B70A85, 0x2E1B2138, 0x4D2C6DFC, 0x53380D13,
0x650A7354, 0x766A0ABB, 0x81C2C92E, 0x92722C85,
0xA2BFE8A1, 0xA81A664B, 0xC24B8B70, 0xC76C51A3,
0xD192E819, 0xD6990624, 0xF40E3585, 0x106AA070,
0x19A4C116, 0x1E376C08, 0x2748774C, 0x34B0BCB5,
0x391C0CB3, 0x4ED8AA4A, 0x5B9CCA4F, 0x682E6FF3,
0x748F82EE, 0x78A5636F, 0x84C87814, 0x8CC70208,
0x90BEFFFA, 0xA4506CEB, 0xBEF9A3F7, 0xC67178F2
};
#define SALSA(a,b,c,d) { \
t = a+d; b^=rotateL(t, 7); \
t = b+a; c^=rotateL(t, 9); \
t = c+b; d^=rotateL(t, 13); \
t = d+c; a^=rotateL(t, 18); \
}
#define SALSA_CORE(state) { \
SALSA(state.s0,state.s4,state.s8,state.sc); \
SALSA(state.s5,state.s9,state.sd,state.s1); \
SALSA(state.sa,state.se,state.s2,state.s6); \
SALSA(state.sf,state.s3,state.s7,state.sb); \
SALSA(state.s0,state.s1,state.s2,state.s3); \
SALSA(state.s5,state.s6,state.s7,state.s4); \
SALSA(state.sa,state.sb,state.s8,state.s9); \
SALSA(state.sf,state.sc,state.sd,state.se); \
}
#if __CUDA_ARCH__ >= 320
static __device__ __forceinline__ uint16 xor_salsa8(const uint16 &Bx)
{
uint32_t t;
uint16 state = Bx;
SALSA_CORE(state);
SALSA_CORE(state);
SALSA_CORE(state);
SALSA_CORE(state);
return(state+Bx);
}
#endif
// sha256
static __device__ __forceinline__ uint32_t bsg2_0(const uint32_t x)
{
uint32_t r1 = ROTR32(x, 2);
uint32_t r2 = ROTR32(x, 13);
uint32_t r3 = ROTR32(x, 22);
return xor3b(r1, r2, r3);
}
static __device__ __forceinline__ uint32_t bsg2_1(const uint32_t x)
{
uint32_t r1 = ROTR32(x, 6);
uint32_t r2 = ROTR32(x, 11);
uint32_t r3 = ROTR32(x, 25);
return xor3b(r1, r2, r3);
}
static __device__ __forceinline__ uint32_t ssg2_0(const uint32_t x)
{
uint64_t r1 = ROTR32(x, 7);
uint64_t r2 = ROTR32(x, 18);
uint64_t r3 = shr_t32(x, 3);
return xor3b(r1, r2, r3);
}
static __device__ __forceinline__ uint32_t ssg2_1(const uint32_t x)
{
uint64_t r1 = ROTR32(x, 17);
uint64_t r2 = ROTR32(x, 19);
uint64_t r3 = shr_t32(x, 10);
return xor3b(r1, r2, r3);
}
static __device__ __forceinline__ void sha2_step1(const uint32_t a, const uint32_t b, const uint32_t c, uint32_t &d, const uint32_t e,
const uint32_t f, const uint32_t g, uint32_t &h, const uint32_t in, const uint32_t Kshared)
{
uint32_t t1, t2;
uint32_t vxandx = xandx(e, f, g);
uint32_t bsg21 = bsg2_1(e);
uint32_t bsg20 = bsg2_0(a);
uint32_t andorv = andor32(a, b, c);
t1 = h + bsg21 + vxandx + Kshared + in;
t2 = bsg20 + andorv;
d = d + t1;
h = t1 + t2;
}
static __device__ __forceinline__ void sha2_step2(const uint32_t a, const uint32_t b, const uint32_t c, uint32_t &d, const uint32_t e,
const uint32_t f, const uint32_t g, uint32_t &h, uint32_t* in, const uint32_t pc, const uint32_t Kshared)
{
uint32_t t1, t2;
int pcidx1 = (pc - 2) & 0xF;
int pcidx2 = (pc - 7) & 0xF;
int pcidx3 = (pc - 15) & 0xF;
uint32_t inx0 = in[pc];
uint32_t inx1 = in[pcidx1];
uint32_t inx2 = in[pcidx2];
uint32_t inx3 = in[pcidx3];
uint32_t ssg21 = ssg2_1(inx1);
uint32_t ssg20 = ssg2_0(inx3);
uint32_t vxandx = xandx(e, f, g);
uint32_t bsg21 = bsg2_1(e);
uint32_t bsg20 = bsg2_0(a);
uint32_t andorv = andor32(a, b, c);
in[pc] = ssg21 + inx2 + ssg20 + inx0;
t1 = h + bsg21 + vxandx + Kshared + in[pc];
t2 = bsg20 + andorv;
d = d + t1;
h = t1 + t2;
}
static __device__ __forceinline__
void sha2_round_body(uint32_t* in, uint32_t* r)
{
uint32_t a = r[0];
uint32_t b = r[1];
uint32_t c = r[2];
uint32_t d = r[3];
uint32_t e = r[4];
uint32_t f = r[5];
uint32_t g = r[6];
uint32_t h = r[7];
sha2_step1(a, b, c, d, e, f, g, h, in[0], Ksha[0]);
sha2_step1(h, a, b, c, d, e, f, g, in[1], Ksha[1]);
sha2_step1(g, h, a, b, c, d, e, f, in[2], Ksha[2]);
sha2_step1(f, g, h, a, b, c, d, e, in[3], Ksha[3]);
sha2_step1(e, f, g, h, a, b, c, d, in[4], Ksha[4]);
sha2_step1(d, e, f, g, h, a, b, c, in[5], Ksha[5]);
sha2_step1(c, d, e, f, g, h, a, b, in[6], Ksha[6]);
sha2_step1(b, c, d, e, f, g, h, a, in[7], Ksha[7]);
sha2_step1(a, b, c, d, e, f, g, h, in[8], Ksha[8]);
sha2_step1(h, a, b, c, d, e, f, g, in[9], Ksha[9]);
sha2_step1(g, h, a, b, c, d, e, f, in[10], Ksha[10]);
sha2_step1(f, g, h, a, b, c, d, e, in[11], Ksha[11]);
sha2_step1(e, f, g, h, a, b, c, d, in[12], Ksha[12]);
sha2_step1(d, e, f, g, h, a, b, c, in[13], Ksha[13]);
sha2_step1(c, d, e, f, g, h, a, b, in[14], Ksha[14]);
sha2_step1(b, c, d, e, f, g, h, a, in[15], Ksha[15]);
#pragma unroll 3
for (int i = 0; i<3; i++) {
sha2_step2(a, b, c, d, e, f, g, h, in, 0, Ksha[16 + 16 * i]);
sha2_step2(h, a, b, c, d, e, f, g, in, 1, Ksha[17 + 16 * i]);
sha2_step2(g, h, a, b, c, d, e, f, in, 2, Ksha[18 + 16 * i]);
sha2_step2(f, g, h, a, b, c, d, e, in, 3, Ksha[19 + 16 * i]);
sha2_step2(e, f, g, h, a, b, c, d, in, 4, Ksha[20 + 16 * i]);
sha2_step2(d, e, f, g, h, a, b, c, in, 5, Ksha[21 + 16 * i]);
sha2_step2(c, d, e, f, g, h, a, b, in, 6, Ksha[22 + 16 * i]);
sha2_step2(b, c, d, e, f, g, h, a, in, 7, Ksha[23 + 16 * i]);
sha2_step2(a, b, c, d, e, f, g, h, in, 8, Ksha[24 + 16 * i]);
sha2_step2(h, a, b, c, d, e, f, g, in, 9, Ksha[25 + 16 * i]);
sha2_step2(g, h, a, b, c, d, e, f, in, 10, Ksha[26 + 16 * i]);
sha2_step2(f, g, h, a, b, c, d, e, in, 11, Ksha[27 + 16 * i]);
sha2_step2(e, f, g, h, a, b, c, d, in, 12, Ksha[28 + 16 * i]);
sha2_step2(d, e, f, g, h, a, b, c, in, 13, Ksha[29 + 16 * i]);
sha2_step2(c, d, e, f, g, h, a, b, in, 14, Ksha[30 + 16 * i]);
sha2_step2(b, c, d, e, f, g, h, a, in, 15, Ksha[31 + 16 * i]);
}
r[0] += a;
r[1] += b;
r[2] += c;
r[3] += d;
r[4] += e;
r[5] += f;
r[6] += g;
r[7] += h;
}
static __device__ __forceinline__ uint8 sha256_64(uint32_t *data)
{
uint32_t __align__(64) in[16];
uint32_t __align__(32) buf[8];
((uint16 *)in)[0] = swapvec((uint16*)data);
((uint8*)buf)[0] = ((uint8*)H256)[0];
sha2_round_body(in, buf);
#pragma unroll 14
for (int i = 0; i<14; i++) { in[i + 1] = 0; }
in[0] = 0x80000000;
in[15] = 0x200;
sha2_round_body(in, buf);
return swapvec((uint8*)buf);
}
static __device__ __forceinline__ uint8 sha256_80(uint32_t nonce)
{
uint32_t __align__(64) in[16];
uint32_t __align__(32) buf[8];
((uint16 *)in)[0] = swapvec((uint16*)c_data);
((uint8*)buf)[0] = ((uint8*)H256)[0];
sha2_round_body(in, buf);
#pragma unroll 3
for (int i = 0; i<3; i++) { in[i] = cuda_swab32(c_data[i + 16]); }
// in[3] = cuda_swab32(nonce);
in[3] = nonce;
in[4] = 0x80000000;
in[15] = 0x280;
#pragma unroll
for (int i = 5; i<15; i++) { in[i] = 0; }
sha2_round_body(in, buf);
return swapvec((uint8*)buf);
}
// Pluck Factor 128
#define SHIFT (1024 * 128)
__global__ __launch_bounds__(256, 1)
void pluck_gpu_hash0_v50(uint32_t threads, uint32_t startNonce)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
#if __CUDA_ARCH__ >= 320
const uint32_t nonce = startNonce + thread;
uint32_t shift = SHIFT * thread;
((uint8*)(hashbuffer + shift))[0] = sha256_80(nonce);
((uint8*)(hashbuffer + shift))[1] = make_uint8(0, 0, 0, 0, 0, 0, 0, 0);
for (int i = 2; i < 5; i++)
{
uint32_t randmax = i * 32 - 4;
uint32_t randseed[16];
uint32_t randbuffer[16];
uint32_t joint[16];
((uint8*)randseed)[0] = __ldg8(&(hashbuffer + shift)[32 * i - 64]);
((uint8*)randseed)[1] = __ldg8(&(hashbuffer + shift)[32 * i - 32]);
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
((uint8*)joint)[0] = ((uint8*)randseed)[1];
#pragma unroll
for (int j = 0; j < 8; j++) {
uint32_t rand = randbuffer[j] % (randmax - 32);
joint[j + 8] = __ldgtoint_unaligned(&(hashbuffer + shift)[rand]);
}
uint8 truc = sha256_64(joint);
((uint8*)(hashbuffer + shift))[i] = truc;
((uint8*)randseed)[0] = ((uint8*)joint)[0];
((uint8*)randseed)[1] = truc;
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
for (int j = 0; j < 32; j += 2)
{
uint32_t rand = randbuffer[j / 2] % randmax;
(hashbuffer + shift)[rand] = __ldg(&(hashbuffer + shift)[randmax + j]);
(hashbuffer + shift)[rand + 1] = __ldg(&(hashbuffer + shift)[randmax + j + 1]);
(hashbuffer + shift)[rand + 2] = __ldg(&(hashbuffer + shift)[randmax + j + 2]);
(hashbuffer + shift)[rand + 3] = __ldg(&(hashbuffer + shift)[randmax + j + 3]);
}
} // main loop
#endif
}
}
__global__ __launch_bounds__(256, 1)
void pluck_gpu_hash_v50(uint32_t threads, uint32_t startNonce, uint32_t *nonceVector)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
#if __CUDA_ARCH__ >= 320
const uint32_t nonce = startNonce + thread;
uint32_t shift = SHIFT * thread;
for (int i = 5; i < HASH_MEMORY - 1; i++)
{
uint32_t randmax = i*32-4;
uint32_t randseed[16];
uint32_t randbuffer[16];
uint32_t joint[16];
uint8 Buffbuffer[2];
((uint8*)randseed)[0] = __ldg8(&(hashbuffer + shift)[32*i-64]);
((uint8*)randseed)[1] = __ldg8(&(hashbuffer + shift)[32*i-32]);
Buffbuffer[0] = __ldg8(&(hashbuffer + shift)[32*i - 128]);
Buffbuffer[1] = __ldg8(&(hashbuffer + shift)[32*i - 96]);
((uint16*)randseed)[0] ^= ((uint16*)Buffbuffer)[0];
((uint16*)randbuffer)[0]= xor_salsa8(((uint16*)randseed)[0]);
((uint8*)joint)[0] = __ldg8(&(hashbuffer + shift)[(i-1)<<5]);
#pragma unroll
for (int j = 0; j < 8; j++) {
uint32_t rand = randbuffer[j] % (randmax - 32);
joint[j+8] = __ldgtoint_unaligned(&(hashbuffer + shift)[rand]);
}
uint8 truc = sha256_64(joint);
((uint8*)(hashbuffer + shift))[i] = truc;
((uint8*)randseed)[0] = ((uint8*)joint)[0];
((uint8*)randseed)[1] = truc;
((uint16*)randseed)[0] ^= ((uint16*)Buffbuffer)[0];
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
for (int j = 0; j < 32; j += 2)
{
uint32_t rand = randbuffer[j / 2] % randmax;
(hashbuffer+shift)[rand] = __ldg(&(hashbuffer+shift)[randmax+j]);
(hashbuffer + shift)[rand + 1] = __ldg(&(hashbuffer + shift)[randmax + j + 1]);
(hashbuffer + shift)[rand + 2] = __ldg(&(hashbuffer + shift)[randmax + j + 2]);
(hashbuffer + shift)[rand + 3] = __ldg(&(hashbuffer + shift)[randmax + j + 3]);
}
} // main loop
uint32_t outbuf = __ldgtoint(&(hashbuffer + shift)[28]);
if (outbuf <= pTarget[7]) {
nonceVector[0] = nonce;
}
#endif
}
}
__global__ __launch_bounds__(128, 3)
void pluck_gpu_hash0(uint32_t threads, uint32_t startNonce)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
#if __CUDA_ARCH__ >= 320
const uint32_t nonce = startNonce + thread;
uint32_t shift = SHIFT * thread;
((uint8*)(hashbuffer + shift))[0] = sha256_80(nonce);
((uint8*)(hashbuffer + shift))[1] = make_uint8(0, 0, 0, 0, 0, 0, 0, 0);
for (int i = 2; i < 5; i++)
{
uint32_t randmax = i * 32 - 4;
uint32_t randseed[16];
uint32_t randbuffer[16];
uint32_t joint[16];
((uint8*)randseed)[0] = __ldg8(&(hashbuffer + shift)[32 * i - 64]);
((uint8*)randseed)[1] = __ldg8(&(hashbuffer + shift)[32 * i - 32]);
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
((uint8*)joint)[0] = ((uint8*)randseed)[1];
#pragma unroll
for (int j = 0; j < 8; j++) {
uint32_t rand = randbuffer[j] % (randmax - 32);
joint[j + 8] = __ldgtoint_unaligned(&(hashbuffer + shift)[rand]);
}
uint8 truc = sha256_64(joint);
((uint8*)(hashbuffer + shift))[i] = truc;
((uint8*)randseed)[0] = ((uint8*)joint)[0];
((uint8*)randseed)[1] = truc;
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
for (int j = 0; j < 32; j += 2)
{
uint32_t rand = randbuffer[j / 2] % randmax;
(hashbuffer + shift)[rand] = __ldg(&(hashbuffer + shift)[randmax + j]);
(hashbuffer + shift)[rand + 1] = __ldg(&(hashbuffer + shift)[randmax + j + 1]);
(hashbuffer + shift)[rand + 2] = __ldg(&(hashbuffer + shift)[randmax + j + 2]);
(hashbuffer + shift)[rand + 3] = __ldg(&(hashbuffer + shift)[randmax + j + 3]);
}
} // main loop
#endif
}
}
__global__ __launch_bounds__(128, 3)
void pluck_gpu_hash(uint32_t threads, uint32_t startNonce, uint32_t *nonceVector)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
#if __CUDA_ARCH__ >= 320
const uint32_t nonce = startNonce + thread;
uint32_t shift = SHIFT * thread;
for (int i = 5; i < HASH_MEMORY - 1; i++)
{
uint32_t randmax = i * 32 - 4;
uint32_t randseed[16];
uint32_t randbuffer[16];
uint32_t joint[16];
uint8 Buffbuffer[2];
((uint8*)randseed)[0] = __ldg8(&(hashbuffer + shift)[32 * i - 64]);
((uint8*)randseed)[1] = __ldg8(&(hashbuffer + shift)[32 * i - 32]);
Buffbuffer[0] = __ldg8(&(hashbuffer + shift)[32 * i - 128]);
Buffbuffer[1] = __ldg8(&(hashbuffer + shift)[32 * i - 96]);
((uint16*)randseed)[0] ^= ((uint16*)Buffbuffer)[0];
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
((uint8*)joint)[0] = __ldg8(&(hashbuffer + shift)[(i - 1) << 5]);
#pragma unroll
for (int j = 0; j < 8; j++)
{
uint32_t rand = randbuffer[j] % (randmax - 32);
joint[j + 8] = __ldgtoint_unaligned(&(hashbuffer + shift)[rand]);
}
uint8 truc = sha256_64(joint);
((uint8*)(hashbuffer + shift))[i] = truc;
((uint8*)randseed)[0] = ((uint8*)joint)[0];
((uint8*)randseed)[1] = truc;
((uint16*)randseed)[0] ^= ((uint16*)Buffbuffer)[0];
((uint16*)randbuffer)[0] = xor_salsa8(((uint16*)randseed)[0]);
for (int j = 0; j < 32; j += 2)
{
uint32_t rand = randbuffer[j / 2] % randmax;
(hashbuffer + shift)[rand] = __ldg(&(hashbuffer + shift)[randmax + j]);
(hashbuffer + shift)[rand + 1] = __ldg(&(hashbuffer + shift)[randmax + j + 1]);
(hashbuffer + shift)[rand + 2] = __ldg(&(hashbuffer + shift)[randmax + j + 2]);
(hashbuffer + shift)[rand + 3] = __ldg(&(hashbuffer + shift)[randmax + j + 3]);
}
} // main loop
uint32_t outbuf = __ldgtoint(&(hashbuffer + shift)[28]);
if (outbuf <= pTarget[7]) {
nonceVector[0] = nonce;
}
#endif
}
}
void pluck_cpu_init(int thr_id, uint32_t threads, uint32_t* hash)
{
cuda_get_arch(thr_id);
cudaMemcpyToSymbol(hashbuffer, &hash, sizeof(hash), 0, cudaMemcpyHostToDevice);
cudaMalloc(&d_PlNonce[thr_id], sizeof(uint32_t));
}
__host__
uint32_t pluck_cpu_hash(int thr_id, uint32_t threads, uint32_t startNounce, int order)
{
uint32_t result[8] = { 0xffffffff };
cudaMemset(d_PlNonce[thr_id], 0xff, sizeof(uint32_t));
const uint32_t threadsperblock = 128;
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
dim3 grid50((threads + 256 - 1) / 256);
dim3 block50(256);
if (device_sm[device_map[thr_id]] <= 300) {
applog(LOG_ERR,"Sorry pluck not supported on SM 3.0 devices");
return 0;
} else if (device_sm[device_map[thr_id]] >= 500) {
pluck_gpu_hash0_v50 <<< grid50, block50 >>>(threads, startNounce);
pluck_gpu_hash_v50 <<< grid50, block50 >>>(threads, startNounce, d_PlNonce[thr_id]);
} else {
pluck_gpu_hash0 <<< grid, block >>>(threads, startNounce);
pluck_gpu_hash <<< grid, block >>>(threads, startNounce, d_PlNonce[thr_id]);
}
//MyStreamSynchronize(NULL, order, thr_id);
CUDA_SAFE_CALL(cudaThreadSynchronize());
cudaMemcpy(&result[thr_id], d_PlNonce[thr_id], sizeof(uint32_t), cudaMemcpyDeviceToHost);
return result[thr_id];
}
__host__
void pluck_setBlockTarget(const void *pdata, const void *ptarget)
{
unsigned char PaddedMessage[80];
memcpy(PaddedMessage, pdata, 80);
cudaMemcpyToSymbol(c_data, PaddedMessage, 80, 0, cudaMemcpyHostToDevice);
cudaMemcpyToSymbol(pTarget, ptarget, 32, 0, cudaMemcpyHostToDevice);
}
+240
View File
@@ -0,0 +1,240 @@
/* Based on djm code */
#include <stdint.h>
#include "miner.h"
#include "cuda_helper.h"
#include <openssl/sha.h>
static uint32_t *d_hash[MAX_GPUS] ;
extern void pluck_setBlockTarget(const void* data, const void *ptarget);
extern void pluck_cpu_init(int thr_id, uint32_t threads, uint32_t *d_outputHash);
extern uint32_t pluck_cpu_hash(int thr_id, uint32_t threads, uint32_t startNounce, int order);
extern float tp_coef[MAX_GPUS];
#define ROTL(a, b) (((a) << (b)) | ((a) >> (32 - (b))))
//note, this is 64 bytes
static inline void xor_salsa8(uint32_t B[16], const uint32_t Bx[16])
{
#define ROTL(a, b) (((a) << (b)) | ((a) >> (32 - (b))))
uint32_t x00, x01, x02, x03, x04, x05, x06, x07, x08, x09, x10, x11, x12, x13, x14, x15;
int i;
x00 = (B[0] ^= Bx[0]);
x01 = (B[1] ^= Bx[1]);
x02 = (B[2] ^= Bx[2]);
x03 = (B[3] ^= Bx[3]);
x04 = (B[4] ^= Bx[4]);
x05 = (B[5] ^= Bx[5]);
x06 = (B[6] ^= Bx[6]);
x07 = (B[7] ^= Bx[7]);
x08 = (B[8] ^= Bx[8]);
x09 = (B[9] ^= Bx[9]);
x10 = (B[10] ^= Bx[10]);
x11 = (B[11] ^= Bx[11]);
x12 = (B[12] ^= Bx[12]);
x13 = (B[13] ^= Bx[13]);
x14 = (B[14] ^= Bx[14]);
x15 = (B[15] ^= Bx[15]);
for (i = 0; i < 8; i += 2) {
/* Operate on columns. */
x04 ^= ROTL(x00 + x12, 7); x09 ^= ROTL(x05 + x01, 7);
x14 ^= ROTL(x10 + x06, 7); x03 ^= ROTL(x15 + x11, 7);
x08 ^= ROTL(x04 + x00, 9); x13 ^= ROTL(x09 + x05, 9);
x02 ^= ROTL(x14 + x10, 9); x07 ^= ROTL(x03 + x15, 9);
x12 ^= ROTL(x08 + x04, 13); x01 ^= ROTL(x13 + x09, 13);
x06 ^= ROTL(x02 + x14, 13); x11 ^= ROTL(x07 + x03, 13);
x00 ^= ROTL(x12 + x08, 18); x05 ^= ROTL(x01 + x13, 18);
x10 ^= ROTL(x06 + x02, 18); x15 ^= ROTL(x11 + x07, 18);
/* Operate on rows. */
x01 ^= ROTL(x00 + x03, 7); x06 ^= ROTL(x05 + x04, 7);
x11 ^= ROTL(x10 + x09, 7); x12 ^= ROTL(x15 + x14, 7);
x02 ^= ROTL(x01 + x00, 9); x07 ^= ROTL(x06 + x05, 9);
x08 ^= ROTL(x11 + x10, 9); x13 ^= ROTL(x12 + x15, 9);
x03 ^= ROTL(x02 + x01, 13); x04 ^= ROTL(x07 + x06, 13);
x09 ^= ROTL(x08 + x11, 13); x14 ^= ROTL(x13 + x12, 13);
x00 ^= ROTL(x03 + x02, 18); x05 ^= ROTL(x04 + x07, 18);
x10 ^= ROTL(x09 + x08, 18); x15 ^= ROTL(x14 + x13, 18);
}
B[0] += x00;
B[1] += x01;
B[2] += x02;
B[3] += x03;
B[4] += x04;
B[5] += x05;
B[6] += x06;
B[7] += x07;
B[8] += x08;
B[9] += x09;
B[10] += x10;
B[11] += x11;
B[12] += x12;
B[13] += x13;
B[14] += x14;
B[15] += x15;
#undef ROTL
}
static void sha256_hash(uchar *hash, const uchar *data, int len)
{
SHA256_CTX ctx;
SHA256_Init(&ctx);
SHA256_Update(&ctx, data, len);
SHA256_Final(hash, &ctx);
}
// hash exactly 64 bytes (ie, sha256 block size)
static void sha256_hash512(uint32_t *hash, const uint32_t *data)
{
uint32_t _ALIGN(64) S[16];
uint32_t _ALIGN(64) T[16];
uchar _ALIGN(64) E[64] = { 0 };
int i;
sha256_init(S);
for (i = 0; i < 16; i++)
T[i] = be32dec(&data[i]);
sha256_transform(S, T, 0);
E[3] = 0x80;
E[61] = 0x02; // T[15] = 8 * 64 => 0x200;
sha256_transform(S, (uint32_t*)E, 0);
for (i = 0; i < 8; i++)
be32enc(&hash[i], S[i]);
}
#define BLOCK_HEADER_SIZE 80
void pluckhash(uint32_t *hash, const uint32_t *data, uchar *hashbuffer, const int N)
{
int size = N * 1024;
sha256_hash(hashbuffer, (uchar*)data, BLOCK_HEADER_SIZE);
memset(&hashbuffer[32], 0, 32);
for (int i = 64; i < size - 32; i += 32)
{
uint32_t _ALIGN(64) randseed[16];
uint32_t _ALIGN(64) randbuffer[16];
uint32_t _ALIGN(64) joint[16];
//i-4 because we use integers for all references against this, and we don't want to go 3 bytes over the defined area
//we could use size here, but then it's probable to use 0 as the value in most cases
int randmax = i - 4;
//setup randbuffer to be an array of random indexes
memcpy(randseed, &hashbuffer[i - 64], 64);
if (i > 128) memcpy(randbuffer, &hashbuffer[i - 128], 64);
else memset(randbuffer, 0, 64);
xor_salsa8((uint32_t*)randbuffer, (uint32_t*)randseed);
memcpy(joint, &hashbuffer[i - 32], 32);
//use the last hash value as the seed
for (int j = 32; j < 64; j += 4)
{
//every other time, change to next random index
//randmax - 32 as otherwise we go beyond memory that's already been written to
uint32_t rand = randbuffer[(j - 32) >> 2] % (randmax - 32);
joint[j >> 2] = *((uint32_t *)&hashbuffer[rand]);
}
sha256_hash512((uint32_t*)&hashbuffer[i], joint);
//setup randbuffer to be an array of random indexes
//use last hash value and previous hash value(post-mixing)
memcpy(randseed, &hashbuffer[i - 32], 64);
if (i > 128) memcpy(randbuffer, &hashbuffer[i - 128], 64);
else memset(randbuffer, 0, 64);
xor_salsa8((uint32_t*)randbuffer, (uint32_t*)randseed);
//use the last hash value as the seed
for (int j = 0; j < 32; j += 2)
{
uint32_t rand = randbuffer[j >> 1] % randmax;
*((uint32_t *)(hashbuffer + rand)) = *((uint32_t *)(hashbuffer + j + randmax));
}
}
memcpy(hash, hashbuffer, 32);
}
static bool init[MAX_GPUS] = { 0 };
static __thread uchar* scratchbuf = NULL;
extern "C" int scanhash_pluck(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
{
const uint32_t first_nonce = pdata[19];
uint32_t endiandata[20];
int opt_pluck_n = 128;
int intensity = is_windows() ? 17 : 19; /* beware > 20 could work and create diff problems later */
uint32_t throughput = device_intensity(thr_id, __func__, 1U << intensity);
// divide by 128 for this algo which require a lot of memory
throughput = throughput / 128 - 256;
throughput = min(throughput, max_nonce - first_nonce + 1);
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x0000ff;
if (!init[thr_id])
{
cudaSetDevice(device_map[thr_id]);
//cudaSetDeviceFlags(cudaDeviceScheduleBlockingSync);
//cudaDeviceSetCacheConfig(cudaFuncCachePreferL1);
cudaMalloc(&d_hash[thr_id], opt_pluck_n * 1024 * throughput);
if (!scratchbuf)
scratchbuf = (uchar*) calloc(opt_pluck_n, 1024);
pluck_cpu_init(thr_id, throughput, d_hash[thr_id]);
CUDA_SAFE_CALL(cudaGetLastError());
applog(LOG_INFO, "Using %d cuda threads", throughput);
init[thr_id] = true;
}
for (int k = 0; k < 20; k++)
be32enc(&endiandata[k], ((uint32_t*)pdata)[k]);
pluck_setBlockTarget(endiandata,ptarget);
do {
uint32_t foundNonce = pluck_cpu_hash(thr_id, throughput, pdata[19], 0);
if (foundNonce != UINT32_MAX)
{
const uint32_t Htarg = ptarget[7];
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
pluckhash(vhash64, endiandata, scratchbuf, opt_pluck_n);
if (vhash64[7] <= Htarg && fulltest(vhash64, ptarget)) {
*hashes_done = pdata[19] - first_nonce + throughput;
pdata[19] = foundNonce;
return 1;
} else {
applog(LOG_WARNING, "GPU #%d: result for %08x does not validate on CPU!", device_map[thr_id], foundNonce);
}
}
pdata[19] += throughput;
} while (pdata[19] < max_nonce && !work_restart[thr_id].restart);
*hashes_done = pdata[19] - first_nonce;
return 0;
}
-361
View File
@@ -1,361 +0,0 @@
/**
* Functions which handle multiple pools data
*/
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include "miner.h"
#include "compat.h"
// to move in miner.h
extern bool allow_gbt;
extern bool allow_mininginfo;
extern bool check_dups;
extern double opt_max_diff;
extern double opt_max_rate;
extern int opt_scantime;
extern int opt_time_limit;
extern char* rpc_url;
extern char* rpc_user;
extern char* rpc_pass;
extern char* short_url;
extern struct work _ALIGN(64) g_work;
extern struct stratum_ctx stratum;
extern pthread_mutex_t stratum_work_lock;
extern bool get_work(struct thr_info *thr, struct work *work);
extern bool stratum_need_reset;
extern time_t firstwork_time;
extern volatile time_t g_work_time;
extern volatile int pool_switch_count;
extern volatile bool pool_is_switching;
extern uint8_t conditional_state[MAX_GPUS];
extern struct option options[];
#define CFG_NULL 0
#define CFG_POOL 1
struct opt_config_array {
int cat;
const char *name; // json key
const char *longname; // global opt name if different
} cfg_array_keys[] = {
{ CFG_POOL, "url", NULL }, /* let this key first, increment pools */
{ CFG_POOL, "user", NULL },
{ CFG_POOL, "pass", NULL },
{ CFG_POOL, "userpass", NULL },
{ CFG_POOL, "name", "pool-name" },
{ CFG_POOL, "scantime", "pool-scantime" },
{ CFG_POOL, "max-diff", "pool-max-diff" },
{ CFG_POOL, "max-rate", "pool-max-rate" },
{ CFG_POOL, "removed", "pool-removed" },
{ CFG_POOL, "disabled", "pool-removed" }, // sample alias
{ CFG_POOL, "time-limit", "pool-time-limit" },
{ CFG_NULL, NULL, NULL }
};
// store current credentials in pools container
void pool_set_creds(int pooln)
{
struct pool_infos *p = &pools[pooln];
snprintf(p->url, sizeof(p->url), "%s", rpc_url);
snprintf(p->short_url, sizeof(p->short_url), "%s", short_url);
snprintf(p->user, sizeof(p->user), "%s", rpc_user);
snprintf(p->pass, sizeof(p->pass), "%s", rpc_pass);
if (!(p->status & POOL_ST_DEFINED)) {
p->id = pooln;
p->status |= POOL_ST_DEFINED;
// init pool options as "unset"
// until cmdline is not fully parsed...
p->max_diff = -1.;
p->max_rate = -1.;
p->scantime = -1;
p->time_limit = -1;
p->allow_mininginfo = allow_mininginfo;
p->allow_gbt = allow_gbt;
p->check_dups = check_dups;
p->status |= POOL_ST_DEFINED;
}
if (strlen(rpc_url)) {
if (!strncasecmp(rpc_url, "stratum", 7))
p->type = POOL_STRATUM;
else /* if (!strncasecmp(rpc_url, "http", 4)) */
p->type = POOL_GETWORK; // todo: or longpoll
p->status |= POOL_ST_VALID;
}
}
// fill the unset pools options with cmdline ones
void pool_init_defaults()
{
struct pool_infos *p;
for (int i=0; i<num_pools; i++) {
p = &pools[i];
if (p->max_diff == -1.) p->max_diff = opt_max_diff;
if (p->max_rate == -1.) p->max_rate = opt_max_rate;
if (p->scantime == -1) p->scantime = opt_scantime;
if (p->time_limit == -1) p->time_limit = opt_time_limit;
}
}
// attributes only set by a json pools config
void pool_set_attr(int pooln, const char* key, char* arg)
{
struct pool_infos *p = &pools[pooln];
if (!strcasecmp(key, "name")) {
snprintf(p->name, sizeof(p->name), "%s", arg);
return;
}
if (!strcasecmp(key, "scantime")) {
p->scantime = atoi(arg);
return;
}
if (!strcasecmp(key, "max-diff")) {
p->max_diff = atof(arg);
return;
}
if (!strcasecmp(key, "max-rate")) {
p->max_rate = atof(arg);
return;
}
if (!strcasecmp(key, "time-limit")) {
p->time_limit = atoi(arg);
return;
}
if (!strcasecmp(key, "removed")) {
int removed = atoi(arg);
if (removed) {
p->status |= POOL_ST_REMOVED;
}
return;
}
}
// pool switching code
bool pool_switch(int pooln)
{
int prevn = cur_pooln;
struct pool_infos *prev = &pools[cur_pooln];
struct pool_infos* p = NULL;
// save prev stratum connection infos (struct)
if (prev->type & POOL_STRATUM) {
// may not be the right moment to free,
// to check if required on submit...
stratum_free_job(&stratum);
prev->stratum = stratum;
}
if (pooln < num_pools) {
cur_pooln = pooln;
p = &pools[cur_pooln];
} else {
applog(LOG_ERR, "Switch to inexistant pool %d!", pooln);
return false;
}
// save global attributes
prev->allow_mininginfo = allow_mininginfo;
prev->allow_gbt = allow_gbt;
prev->check_dups = check_dups;
pthread_mutex_lock(&stratum_work_lock);
free(rpc_user); rpc_user = strdup(p->user);
free(rpc_pass); rpc_pass = strdup(p->pass);
free(rpc_url); rpc_url = strdup(p->url);
short_url = p->short_url; // just a pointer, no alloc
opt_scantime = p->scantime;
opt_max_diff = p->max_diff;
opt_max_rate = p->max_rate;
opt_time_limit = p->time_limit;
want_stratum = have_stratum = (p->type & POOL_STRATUM) != 0;
pthread_mutex_unlock(&stratum_work_lock);
if (prevn != cur_pooln) {
pool_switch_count++;
g_work_time = 0;
g_work.data[0] = 0;
pool_is_switching = true;
stratum_need_reset = true;
// used to get the pool uptime
firstwork_time = time(NULL);
restart_threads();
// reset wait states
for (int n=0; n<opt_n_threads; n++)
conditional_state[n] = false;
// restore flags
allow_gbt = p->allow_gbt;
allow_mininginfo = p->allow_mininginfo;
check_dups = p->check_dups;
if (want_stratum) {
// temporary... until stratum code cleanup
stratum = p->stratum;
stratum.pooln = cur_pooln;
// unlock the stratum thread
tq_push(thr_info[stratum_thr_id].q, strdup(rpc_url));
applog(LOG_BLUE, "Switch to stratum pool %d: %s", cur_pooln,
strlen(p->name) ? p->name : p->short_url);
} else {
applog(LOG_BLUE, "Switch to pool %d: %s", cur_pooln,
strlen(p->name) ? p->name : p->short_url);
}
// will unlock the longpoll thread on /LP url receive
want_longpoll = (p->type & POOL_LONGPOLL) || !(p->type & POOL_STRATUM);
if (want_longpoll) {
pthread_mutex_lock(&stratum_work_lock);
// will issue a lp_url request to unlock the longpoll thread
have_longpoll = false;
get_work(&thr_info[0], &g_work);
pthread_mutex_unlock(&stratum_work_lock);
}
}
return true;
}
// search available pool
int pool_get_first_valid(int startfrom)
{
int next = 0;
struct pool_infos *p;
for (int i=0; i<num_pools; i++) {
int pooln = (startfrom + i) % num_pools;
p = &pools[pooln];
if (!(p->status & POOL_ST_VALID))
continue;
if (p->status & (POOL_ST_DISABLED | POOL_ST_REMOVED))
continue;
next = pooln;
break;
}
return next;
}
// switch to next available pool
bool pool_switch_next()
{
if (num_pools > 1) {
int pooln = pool_get_first_valid(cur_pooln+1);
return pool_switch(pooln);
} else {
// no switch possible
if (!opt_quiet)
applog(LOG_DEBUG, "No other pools to try...");
return false;
}
}
// seturl from api remote (deprecated)
bool pool_switch_url(char *params)
{
int prevn = cur_pooln, nextn;
parse_arg('o', params);
// cur_pooln modified by parse_arg('o'), get new pool num
nextn = cur_pooln;
// and to handle the "hot swap" from current one...
cur_pooln = prevn;
if (nextn == prevn)
return false;
return pool_switch(nextn);
}
// Parse pools array in json config
bool parse_pool_array(json_t *obj)
{
size_t idx;
json_t *p, *val;
if (!json_is_array(obj))
return false;
// array of objects [ {}, {} ]
json_array_foreach(obj, idx, p)
{
if (!json_is_object(p))
continue;
for (int i = 0; i < ARRAY_SIZE(cfg_array_keys); i++)
{
int opt = -1;
char *s = NULL;
if (cfg_array_keys[i].cat != CFG_POOL)
continue;
val = json_object_get(p, cfg_array_keys[i].name);
if (!val)
continue;
for (int k = 0; k < options_count(); k++)
{
const char *alias = cfg_array_keys[i].longname;
if (alias && !strcasecmp(options[k].name, alias)) {
opt = k;
break;
}
if (!alias && !strcasecmp(options[k].name, cfg_array_keys[i].name)) {
opt = k;
break;
}
}
if (opt == -1)
continue;
if (json_is_string(val)) {
s = strdup(json_string_value(val));
if (!s)
continue;
// applog(LOG_DEBUG, "pool key %s '%s'", options[opt].name, s);
parse_arg(options[opt].val, s);
free(s);
} else {
// numeric or bool
char buf[32] = { 0 };
double d = 0.;
if (json_is_true(val)) d = 1.;
else if (json_is_integer(val))
d = 1.0 * json_integer_value(val);
else if (json_is_real(val))
d = json_real_value(val);
snprintf(buf, sizeof(buf)-1, "%f", d);
// applog(LOG_DEBUG, "pool key %s '%f'", options[opt].name, d);
parse_arg(options[opt].val, buf);
}
}
}
return true;
}
// debug stuff
void pool_dump_infos()
{
struct pool_infos *p;
for (int i=0; i<num_pools; i++) {
p = &pools[i];
applog(LOG_DEBUG, "POOL %01d: %s USER %s -s %d", i,
p->short_url, p->user, p->scantime);
}
}
+193 -191
View File
@@ -11,15 +11,20 @@ __constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + paddi
#undef SHR
#undef CONST_EXP2
//#define SHL(x, n) ((x) << (n))
//#define SHR(x, n) ((x) >> (n))
#define SHR(x, n) SHR2(x, n)
#define SHL(x, n) SHL2(x, n)
#define ROL(x, n) ROL2(x, n)
#undef ROTL64
#define ROTL64 ROL2
#define CONST_EXP2(i) \
q[i+0] + ROL(q[i+1], 5) + q[i+2] + ROL(q[i+3], 11) + \
q[i+4] + ROL(q[i+5], 27) + q[i+6] + SWAPUINT2(q[i+7]) + \
q[i+8] + ROL(q[i+9], 37) + q[i+10] + ROL(q[i+11], 43) + \
q[i+12] + ROL(q[i+13], 53) + (SHR(q[i+14],1) ^ q[i+14]) + (SHR(q[i+15],2) ^ q[i+15])
q[i+0] + ROTL64(q[i+1], 5) + q[i+2] + ROTL64(q[i+3], 11) + \
q[i+4] + ROTL64(q[i+5], 27) + q[i+6] + SWAPUINT2(q[i+7]) + \
q[i+8] + ROTL64(q[i+9], 37) + q[i+10] + ROTL64(q[i+11], 43) + \
q[i+12] + ROTL64(q[i+13], 53) + (SHR(q[i+14],1) ^ q[i+14]) + (SHR(q[i+15],2) ^ q[i+15])
__device__ void Compression512_64_first(uint2 *msg, uint2 *hash)
{
@@ -28,138 +33,136 @@ __device__ void Compression512_64_first(uint2 *msg, uint2 *hash)
uint2 tmp;
tmp = (msg[5] ^ hash[5]) - (msg[7] ^ hash[7]) + (hash[10]) + (hash[13]) + (hash[14]);
q[0] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[1];
q[0] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[1];
tmp = (msg[6] ^ hash[6]) - (msg[8] ^ hash[8]) + (hash[11]) + (hash[14]) - (msg[15] ^ hash[15]);
q[1] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[2];
q[1] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[2];
tmp = (msg[0] ^ hash[0]) + (msg[7] ^ hash[7]) + (hash[9]) - (hash[12]) + (msg[15] ^ hash[15]);
q[2] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[3];
q[2] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[3];
tmp = (msg[0] ^ hash[0]) - (msg[1] ^ hash[1]) + (msg[8] ^ hash[8]) - (hash[10]) + (hash[13]);
q[3] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[4];
q[3] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[4];
tmp = (msg[1] ^ hash[1]) + (msg[2] ^ hash[2]) + (hash[9]) - (hash[11]) - (hash[14]);
q[4] = (SHR(tmp, 1) ^ tmp) + hash[5];
tmp = (msg[3] ^ hash[3]) - (msg[2] ^ hash[2]) + (hash[10]) - (hash[12]) + (msg[15] ^ hash[15]);
q[5] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[6];
q[5] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[6];
tmp = (msg[4] ^ hash[4]) - (msg[0] ^ hash[0]) - (msg[3] ^ hash[3]) - (hash[11]) + (hash[13]);
q[6] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[7];
q[6] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[7];
tmp = (msg[1] ^ hash[1]) - (msg[4] ^ hash[4]) - (msg[5] ^ hash[5]) - (hash[12]) - (hash[14]);
q[7] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[8];
q[7] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[8];
tmp = (msg[2] ^ hash[2]) - (msg[5] ^ hash[5]) - (msg[6] ^ hash[6]) + (hash[13]) - (msg[15] ^ hash[15]);
q[8] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[9];
q[8] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[9];
tmp = (msg[0] ^ hash[0]) - (msg[3] ^ hash[3]) + (msg[6] ^ hash[6]) - (msg[7] ^ hash[7]) + (hash[14]);
q[9] = (SHR(tmp, 1) ^ tmp) + hash[10];
tmp = (msg[8] ^ hash[8]) - (msg[1] ^ hash[1]) - (msg[4] ^ hash[4]) - (msg[7] ^ hash[7]) + (msg[15] ^ hash[15]);
q[10] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[11];
q[10] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[11];
tmp = (msg[8] ^ hash[8]) - (msg[0] ^ hash[0]) - (msg[2] ^ hash[2]) - (msg[5] ^ hash[5]) + (hash[9]);
q[11] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[12];
q[11] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[12];
tmp = (msg[1] ^ hash[1]) + (msg[3] ^ hash[3]) - (msg[6] ^ hash[6]) - (hash[9]) + (hash[10]);
q[12] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[13];
q[12] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[13];
tmp = (msg[2] ^ hash[2]) + (msg[4] ^ hash[4]) + (msg[7] ^ hash[7]) + (hash[10]) + (hash[11]);
q[13] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[14];
q[13] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[14];
tmp = (msg[3] ^ hash[3]) - (msg[5] ^ hash[5]) + (msg[8] ^ hash[8]) - (hash[11]) - (hash[12]);
q[14] = (SHR(tmp, 1) ^ tmp) + hash[15];
tmp = (msg[12] ^ hash[12]) - (msg[4] ^ hash[4]) - (msg[6] ^ hash[6]) - (hash[9]) + (hash[13]);
q[15] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[0];
q[15] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[0];
q[0 + 16] =
(SHR(q[0], 1) ^ SHL(q[0], 2) ^ ROL(q[0], 13) ^ ROL(q[0], 43)) +
(SHR(q[0 + 1], 2) ^ SHL(q[0 + 1], 1) ^ ROL(q[0 + 1], 19) ^ ROL(q[0 + 1], 53)) +
(SHR(q[0 + 2], 2) ^ SHL(q[0 + 2], 2) ^ ROL(q[0 + 2], 28) ^ ROL(q[0 + 2], 59)) +
(SHR(q[0 + 3], 1) ^ SHL(q[0 + 3], 3) ^ ROL(q[0 + 3], 4) ^ ROL(q[0 + 3], 37)) +
(SHR(q[0 + 4], 1) ^ SHL(q[0 + 4], 2) ^ ROL(q[0 + 4], 13) ^ ROL(q[0 + 4], 43)) +
(SHR(q[0 + 5], 2) ^ SHL(q[0 + 5], 1) ^ ROL(q[0 + 5], 19) ^ ROL(q[0 + 5], 53)) +
(SHR(q[0 + 6], 2) ^ SHL(q[0 + 6], 2) ^ ROL(q[0 + 6], 28) ^ ROL(q[0 + 6], 59)) +
(SHR(q[0 + 7], 1) ^ SHL(q[0 + 7], 3) ^ ROL(q[0 + 7], 4) ^ ROL(q[0 + 7], 37)) +
(SHR(q[0 + 8], 1) ^ SHL(q[0 + 8], 2) ^ ROL(q[0 + 8], 13) ^ ROL(q[0 + 8], 43)) +
(SHR(q[0 + 9], 2) ^ SHL(q[0 + 9], 1) ^ ROL(q[0 + 9], 19) ^ ROL(q[0 + 9], 53)) +
(SHR(q[0 + 10], 2) ^ SHL(q[0 + 10], 2) ^ ROL(q[0 + 10], 28) ^ ROL(q[0 + 10], 59)) +
(SHR(q[0 + 11], 1) ^ SHL(q[0 + 11], 3) ^ ROL(q[0 + 11], 4) ^ ROL(q[0 + 11], 37)) +
(SHR(q[0 + 12], 1) ^ SHL(q[0 + 12], 2) ^ ROL(q[0 + 12], 13) ^ ROL(q[0 + 12], 43)) +
(SHR(q[0 + 13], 2) ^ SHL(q[0 + 13], 1) ^ ROL(q[0 + 13], 19) ^ ROL(q[0 + 13], 53)) +
(SHR(q[0 + 14], 2) ^ SHL(q[0 + 14], 2) ^ ROL(q[0 + 14], 28) ^ ROL(q[0 + 14], 59)) +
(SHR(q[0 + 15], 1) ^ SHL(q[0 + 15], 3) ^ ROL(q[0 + 15], 4) ^ ROL(q[0 + 15], 37)) +
((make_uint2(0x55555550ul,0x55555555) + ROL(msg[0], 0 + 1) +
ROL(msg[0 + 3], 0 + 4)) ^ hash[0 + 7]);
(SHR(q[0], 1) ^ SHL(q[0], 2) ^ ROTL64(q[0], 13) ^ ROTL64(q[0], 43)) +
(SHR(q[0 + 1], 2) ^ SHL(q[0 + 1], 1) ^ ROTL64(q[0 + 1], 19) ^ ROTL64(q[0 + 1], 53)) +
(SHR(q[0 + 2], 2) ^ SHL(q[0 + 2], 2) ^ ROTL64(q[0 + 2], 28) ^ ROTL64(q[0 + 2], 59)) +
(SHR(q[0 + 3], 1) ^ SHL(q[0 + 3], 3) ^ ROTL64(q[0 + 3], 4) ^ ROTL64(q[0 + 3], 37)) +
(SHR(q[0 + 4], 1) ^ SHL(q[0 + 4], 2) ^ ROTL64(q[0 + 4], 13) ^ ROTL64(q[0 + 4], 43)) +
(SHR(q[0 + 5], 2) ^ SHL(q[0 + 5], 1) ^ ROTL64(q[0 + 5], 19) ^ ROTL64(q[0 + 5], 53)) +
(SHR(q[0 + 6], 2) ^ SHL(q[0 + 6], 2) ^ ROTL64(q[0 + 6], 28) ^ ROTL64(q[0 + 6], 59)) +
(SHR(q[0 + 7], 1) ^ SHL(q[0 + 7], 3) ^ ROTL64(q[0 + 7], 4) ^ ROTL64(q[0 + 7], 37)) +
(SHR(q[0 + 8], 1) ^ SHL(q[0 + 8], 2) ^ ROTL64(q[0 + 8], 13) ^ ROTL64(q[0 + 8], 43)) +
(SHR(q[0 + 9], 2) ^ SHL(q[0 + 9], 1) ^ ROTL64(q[0 + 9], 19) ^ ROTL64(q[0 + 9], 53)) +
(SHR(q[0 + 10], 2) ^ SHL(q[0 + 10], 2) ^ ROTL64(q[0 + 10], 28) ^ ROTL64(q[0 + 10], 59)) +
(SHR(q[0 + 11], 1) ^ SHL(q[0 + 11], 3) ^ ROTL64(q[0 + 11], 4) ^ ROTL64(q[0 + 11], 37)) +
(SHR(q[0 + 12], 1) ^ SHL(q[0 + 12], 2) ^ ROTL64(q[0 + 12], 13) ^ ROTL64(q[0 + 12], 43)) +
(SHR(q[0 + 13], 2) ^ SHL(q[0 + 13], 1) ^ ROTL64(q[0 + 13], 19) ^ ROTL64(q[0 + 13], 53)) +
(SHR(q[0 + 14], 2) ^ SHL(q[0 + 14], 2) ^ ROTL64(q[0 + 14], 28) ^ ROTL64(q[0 + 14], 59)) +
(SHR(q[0 + 15], 1) ^ SHL(q[0 + 15], 3) ^ ROTL64(q[0 + 15], 4) ^ ROTL64(q[0 + 15], 37)) +
((make_uint2(0x55555550ul,0x55555555) + ROTL64(msg[0], 0 + 1) +
ROTL64(msg[0 + 3], 0 + 4)) ^ hash[0 + 7]);
q[1 + 16] =
(SHR(q[1], 1) ^ SHL(q[1], 2) ^ ROL(q[1], 13) ^ ROL(q[1], 43)) +
(SHR(q[1 + 1], 2) ^ SHL(q[1 + 1], 1) ^ ROL(q[1 + 1], 19) ^ ROL(q[1 + 1], 53)) +
(SHR(q[1 + 2], 2) ^ SHL(q[1 + 2], 2) ^ ROL(q[1 + 2], 28) ^ ROL(q[1 + 2], 59)) +
(SHR(q[1 + 3], 1) ^ SHL(q[1 + 3], 3) ^ ROL(q[1 + 3], 4) ^ ROL(q[1 + 3], 37)) +
(SHR(q[1 + 4], 1) ^ SHL(q[1 + 4], 2) ^ ROL(q[1 + 4], 13) ^ ROL(q[1 + 4], 43)) +
(SHR(q[1 + 5], 2) ^ SHL(q[1 + 5], 1) ^ ROL(q[1 + 5], 19) ^ ROL(q[1 + 5], 53)) +
(SHR(q[1 + 6], 2) ^ SHL(q[1 + 6], 2) ^ ROL(q[1 + 6], 28) ^ ROL(q[1 + 6], 59)) +
(SHR(q[1 + 7], 1) ^ SHL(q[1 + 7], 3) ^ ROL(q[1 + 7], 4) ^ ROL(q[1 + 7], 37)) +
(SHR(q[1 + 8], 1) ^ SHL(q[1 + 8], 2) ^ ROL(q[1 + 8], 13) ^ ROL(q[1 + 8], 43)) +
(SHR(q[1 + 9], 2) ^ SHL(q[1 + 9], 1) ^ ROL(q[1 + 9], 19) ^ ROL(q[1 + 9], 53)) +
(SHR(q[1 + 10], 2) ^ SHL(q[1 + 10], 2) ^ ROL(q[1 + 10], 28) ^ ROL(q[1 + 10], 59)) +
(SHR(q[1 + 11], 1) ^ SHL(q[1 + 11], 3) ^ ROL(q[1 + 11], 4) ^ ROL(q[1 + 11], 37)) +
(SHR(q[1 + 12], 1) ^ SHL(q[1 + 12], 2) ^ ROL(q[1 + 12], 13) ^ ROL(q[1 + 12], 43)) +
(SHR(q[1 + 13], 2) ^ SHL(q[1 + 13], 1) ^ ROL(q[1 + 13], 19) ^ ROL(q[1 + 13], 53)) +
(SHR(q[1 + 14], 2) ^ SHL(q[1 + 14], 2) ^ ROL(q[1 + 14], 28) ^ ROL(q[1 + 14], 59)) +
(SHR(q[1 + 15], 1) ^ SHL(q[1 + 15], 3) ^ ROL(q[1 + 15], 4) ^ ROL(q[1 + 15], 37)) +
((make_uint2(0xAAAAAAA5, 0x5AAAAAAA) + ROL(msg[1], 1 + 1) +
ROL(msg[1 + 3], 1 + 4)) ^ hash[1 + 7]);
(SHR(q[1], 1) ^ SHL(q[1], 2) ^ ROTL64(q[1], 13) ^ ROTL64(q[1], 43)) +
(SHR(q[1 + 1], 2) ^ SHL(q[1 + 1], 1) ^ ROTL64(q[1 + 1], 19) ^ ROTL64(q[1 + 1], 53)) +
(SHR(q[1 + 2], 2) ^ SHL(q[1 + 2], 2) ^ ROTL64(q[1 + 2], 28) ^ ROTL64(q[1 + 2], 59)) +
(SHR(q[1 + 3], 1) ^ SHL(q[1 + 3], 3) ^ ROTL64(q[1 + 3], 4) ^ ROTL64(q[1 + 3], 37)) +
(SHR(q[1 + 4], 1) ^ SHL(q[1 + 4], 2) ^ ROTL64(q[1 + 4], 13) ^ ROTL64(q[1 + 4], 43)) +
(SHR(q[1 + 5], 2) ^ SHL(q[1 + 5], 1) ^ ROTL64(q[1 + 5], 19) ^ ROTL64(q[1 + 5], 53)) +
(SHR(q[1 + 6], 2) ^ SHL(q[1 + 6], 2) ^ ROTL64(q[1 + 6], 28) ^ ROTL64(q[1 + 6], 59)) +
(SHR(q[1 + 7], 1) ^ SHL(q[1 + 7], 3) ^ ROTL64(q[1 + 7], 4) ^ ROTL64(q[1 + 7], 37)) +
(SHR(q[1 + 8], 1) ^ SHL(q[1 + 8], 2) ^ ROTL64(q[1 + 8], 13) ^ ROTL64(q[1 + 8], 43)) +
(SHR(q[1 + 9], 2) ^ SHL(q[1 + 9], 1) ^ ROTL64(q[1 + 9], 19) ^ ROTL64(q[1 + 9], 53)) +
(SHR(q[1 + 10], 2) ^ SHL(q[1 + 10], 2) ^ ROTL64(q[1 + 10], 28) ^ ROTL64(q[1 + 10], 59)) +
(SHR(q[1 + 11], 1) ^ SHL(q[1 + 11], 3) ^ ROTL64(q[1 + 11], 4) ^ ROTL64(q[1 + 11], 37)) +
(SHR(q[1 + 12], 1) ^ SHL(q[1 + 12], 2) ^ ROTL64(q[1 + 12], 13) ^ ROTL64(q[1 + 12], 43)) +
(SHR(q[1 + 13], 2) ^ SHL(q[1 + 13], 1) ^ ROTL64(q[1 + 13], 19) ^ ROTL64(q[1 + 13], 53)) +
(SHR(q[1 + 14], 2) ^ SHL(q[1 + 14], 2) ^ ROTL64(q[1 + 14], 28) ^ ROTL64(q[1 + 14], 59)) +
(SHR(q[1 + 15], 1) ^ SHL(q[1 + 15], 3) ^ ROTL64(q[1 + 15], 4) ^ ROTL64(q[1 + 15], 37)) +
((make_uint2(0xAAAAAAA5, 0x5AAAAAAA) + ROTL64(msg[1], 1 + 1) +
ROTL64(msg[1 + 3], 1 + 4)) ^ hash[1 + 7]);
q[2 + 16] = CONST_EXP2(2) +
((make_uint2(0xFFFFFFFA, 0x5FFFFFFF) + ROL(msg[2], 2 + 1) +
ROL(msg[2 + 3], 2 + 4) - ROL(msg[2 + 10], 2 + 11)) ^ hash[2 + 7]);
((make_uint2(0xFFFFFFFA, 0x5FFFFFFF) + ROTL64(msg[2], 2 + 1) +
ROTL64(msg[2 + 3], 2 + 4) - ROTL64(msg[2 + 10], 2 + 11)) ^ hash[2 + 7]);
q[3 + 16] = CONST_EXP2(3) +
((make_uint2(0x5555554F, 0x65555555) + ROL(msg[3], 3 + 1) +
ROL(msg[3 + 3], 3 + 4) - ROL(msg[3 + 10], 3 + 11)) ^ hash[3 + 7]);
((make_uint2(0x5555554F, 0x65555555) + ROTL64(msg[3], 3 + 1) +
ROTL64(msg[3 + 3], 3 + 4) - ROTL64(msg[3 + 10], 3 + 11)) ^ hash[3 + 7]);
q[4 + 16] = CONST_EXP2(4) +
((make_uint2(0xAAAAAAA4, 0x6AAAAAAA) +ROL(msg[4], 4 + 1) +
ROL(msg[4 + 3], 4 + 4) - ROL(msg[4 + 10], 4 + 11)) ^ hash[4 + 7]);
((make_uint2(0xAAAAAAA4, 0x6AAAAAAA) +ROTL64(msg[4], 4 + 1) +
ROTL64(msg[4 + 3], 4 + 4) - ROTL64(msg[4 + 10], 4 + 11)) ^ hash[4 + 7]);
q[5 + 16] = CONST_EXP2(5) +
((make_uint2(0xFFFFFFF9, 0x6FFFFFFF) + ROL(msg[5], 5 + 1) +
ROL(msg[5 + 3], 5 + 4) - ROL(msg[5 + 10], 5 + 11)) ^ hash[5 + 7]);
((make_uint2(0xFFFFFFF9, 0x6FFFFFFF) + ROTL64(msg[5], 5 + 1) +
ROTL64(msg[5 + 3], 5 + 4) - ROTL64(msg[5 + 10], 5 + 11)) ^ hash[5 + 7]);
#pragma unroll 3
for (int i = 6; i<9; i++) {
q[i + 16] = CONST_EXP2(i) +
((vectorize((i + 16)*(0x0555555555555555ull)) + ROL(msg[i], i + 1) -
ROL(msg[i - 6], (i - 6) + 1)) ^ hash[i + 7]);
((vectorize((i + 16)*(0x0555555555555555ull)) + ROTL64(msg[i], i + 1) -
ROTL64(msg[i - 6], (i - 6) + 1)) ^ hash[i + 7]);
}
#pragma unroll 4
for (int i = 9; i<13; i++) {
q[i + 16] = CONST_EXP2(i) +
((vectorize((i + 16)*(0x0555555555555555ull)) +
ROL(msg[i + 3], i + 4) - ROL(msg[i - 6], (i - 6) + 1)) ^ hash[i - 9]);
ROTL64(msg[i + 3], i + 4) - ROTL64(msg[i - 6], (i - 6) + 1)) ^ hash[i - 9]);
}
q[13 + 16] = CONST_EXP2(13) +
((make_uint2(0xAAAAAAA1, 0x9AAAAAAA) + ROL(msg[13], 13 + 1) +
ROL(msg[13 - 13], (13 - 13) + 1) - ROL(msg[13 - 6], (13 - 6) + 1)) ^ hash[13 - 9]);
((make_uint2(0xAAAAAAA1, 0x9AAAAAAA) + ROTL64(msg[13], 13 + 1) +
ROTL64(msg[13 - 13], (13 - 13) + 1) - ROTL64(msg[13 - 6], (13 - 6) + 1)) ^ hash[13 - 9]);
q[14 + 16] = CONST_EXP2(14) +
((make_uint2(0xFFFFFFF6, 0x9FFFFFFF) + ROL(msg[14], 14 + 1) +
ROL(msg[14 - 13], (14 - 13) + 1) - ROL(msg[14 - 6], (14 - 6) + 1)) ^ hash[14 - 9]);
((make_uint2(0xFFFFFFF6, 0x9FFFFFFF) + ROTL64(msg[14], 14 + 1) +
ROTL64(msg[14 - 13], (14 - 13) + 1) - ROTL64(msg[14 - 6], (14 - 6) + 1)) ^ hash[14 - 9]);
q[15 + 16] = CONST_EXP2(15) +
((make_uint2(0x5555554B, 0xA5555555) + ROL(msg[15], 15 + 1) +
ROL(msg[15 - 13], (15 - 13) + 1) - ROL(msg[15 - 6], (15 - 6) + 1)) ^ hash[15 - 9]);
((make_uint2(0x5555554B, 0xA5555555) + ROTL64(msg[15], 15 + 1) +
ROTL64(msg[15 - 13], (15 - 13) + 1) - ROTL64(msg[15 - 6], (15 - 6) + 1)) ^ hash[15 - 9]);
uint2 XL64 = q[16] ^ q[17] ^ q[18] ^ q[19] ^ q[20] ^ q[21] ^ q[22] ^ q[23];
uint2 XH64 = XL64^q[24] ^ q[25] ^ q[26] ^ q[27] ^ q[28] ^ q[29] ^ q[30] ^ q[31];
hash[0] = (SHL(XH64, 5) ^ SHR(q[16], 5) ^ msg[0]) + (XL64 ^ q[24] ^ q[0]);
hash[1] = (SHR(XH64, 7) ^ SHL(q[17], 8) ^ msg[1]) + (XL64 ^ q[25] ^ q[1]);
hash[2] = (SHR(XH64, 5) ^ SHL(q[18], 5) ^ msg[2]) + (XL64 ^ q[26] ^ q[2]);
hash[3] = (SHR(XH64, 1) ^ SHL(q[19], 5) ^ msg[3]) + (XL64 ^ q[27] ^ q[3]);
hash[4] = (SHR(XH64, 3) ^ q[20] ^ msg[4]) + (XL64 ^ q[28] ^ q[4]);
hash[5] = (SHL(XH64, 6) ^ SHR(q[21], 6) ^ msg[5]) + (XL64 ^ q[29] ^ q[5]);
hash[6] = (SHR(XH64, 4) ^ SHL(q[22], 6) ^ msg[6]) + (XL64 ^ q[30] ^ q[6]);
hash[7] = (SHR(XH64, 11) ^ SHL(q[23], 2) ^ msg[7]) + (XL64 ^ q[31] ^ q[7]);
hash[0] = (SHL(XH64, 5) ^ SHR(q[16], 5) ^ msg[0]) + (XL64 ^ q[24] ^ q[0]);
hash[1] = (SHR(XH64, 7) ^ SHL(q[17], 8) ^ msg[1]) + (XL64 ^ q[25] ^ q[1]);
hash[2] = (SHR(XH64, 5) ^ SHL(q[18], 5) ^ msg[2]) + (XL64 ^ q[26] ^ q[2]);
hash[3] = (SHR(XH64, 1) ^ SHL(q[19], 5) ^ msg[3]) + (XL64 ^ q[27] ^ q[3]);
hash[4] = (SHR(XH64, 3) ^ q[20] ^ msg[4]) + (XL64 ^ q[28] ^ q[4]);
hash[5] = (SHL(XH64, 6) ^ SHR(q[21], 6) ^ msg[5]) + (XL64 ^ q[29] ^ q[5]);
hash[6] = (SHR(XH64, 4) ^ SHL(q[22], 6) ^ msg[6]) + (XL64 ^ q[30] ^ q[6]);
hash[7] = (SHR(XH64, 11) ^ SHL(q[23], 2) ^ msg[7]) + (XL64 ^ q[31] ^ q[7]);
hash[8] = ROL(hash[4], 9) + (XH64 ^ q[24] ^ msg[8]) + (SHL(XL64, 8) ^ q[23] ^ q[8]);
hash[9] = ROL(hash[5], 10) + (XH64 ^ q[25]) + (SHR(XL64, 6) ^ q[16] ^ q[9]);
hash[10] = ROL(hash[6], 11) + (XH64 ^ q[26]) + (SHL(XL64, 6) ^ q[17] ^ q[10]);
hash[11] = ROL(hash[7], 12) + (XH64 ^ q[27]) + (SHL(XL64, 4) ^ q[18] ^ q[11]);
hash[12] = ROL(hash[0], 13) + (XH64 ^ q[28]) + (SHR(XL64, 3) ^ q[19] ^ q[12]);
hash[13] = ROL(hash[1], 14) + (XH64 ^ q[29]) + (SHR(XL64, 4) ^ q[20] ^ q[13]);
hash[14] = ROL(hash[2], 15) + (XH64 ^ q[30]) + (SHR(XL64, 7) ^ q[21] ^ q[14]);
hash[15] = ROL(hash[3], 16) + (XH64 ^ q[31] ^ msg[15]) + (SHR(XL64, 2) ^ q[22] ^ q[15]);
hash[8] = ROTL64(hash[4], 9) + (XH64 ^ q[24] ^ msg[8]) + (SHL(XL64, 8) ^ q[23] ^ q[8]);
hash[9] = ROTL64(hash[5], 10) + (XH64 ^ q[25]) + (SHR(XL64, 6) ^ q[16] ^ q[9]);
hash[10] = ROTL64(hash[6], 11) + (XH64 ^ q[26]) + (SHL(XL64, 6) ^ q[17] ^ q[10]);
hash[11] = ROTL64(hash[7], 12) + (XH64 ^ q[27]) + (SHL(XL64, 4) ^ q[18] ^ q[11]);
hash[12] = ROTL64(hash[0], 13) + (XH64 ^ q[28]) + (SHR(XL64, 3) ^ q[19] ^ q[12]);
hash[13] = ROTL64(hash[1], 14) + (XH64 ^ q[29]) + (SHR(XL64, 4) ^ q[20] ^ q[13]);
hash[14] = ROTL64(hash[2], 15) + (XH64 ^ q[30]) + (SHR(XL64, 7) ^ q[21] ^ q[14]);
hash[15] = ROTL64(hash[3], 16) + (XH64 ^ q[31] ^ msg[15]) + (SHR(XL64, 2) ^ q[22] ^ q[15]);
}
__device__
@@ -170,141 +173,141 @@ void Compression512(uint2 *msg, uint2 *hash)
uint2 tmp;
tmp = (msg[ 5] ^ hash[ 5]) - (msg[ 7] ^ hash[ 7]) + (msg[10] ^ hash[10]) + (msg[13] ^ hash[13]) + (msg[14] ^ hash[14]);
q[0] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[1];
q[0] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[1];
tmp = (msg[ 6] ^ hash[ 6]) - (msg[ 8] ^ hash[ 8]) + (msg[11] ^ hash[11]) + (msg[14] ^ hash[14]) - (msg[15] ^ hash[15]);
q[1] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[2];
q[1] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[2];
tmp = (msg[ 0] ^ hash[ 0]) + (msg[ 7] ^ hash[ 7]) + (msg[ 9] ^ hash[ 9]) - (msg[12] ^ hash[12]) + (msg[15] ^ hash[15]);
q[2] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[3];
q[2] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[3];
tmp = (msg[ 0] ^ hash[ 0]) - (msg[ 1] ^ hash[ 1]) + (msg[ 8] ^ hash[ 8]) - (msg[10] ^ hash[10]) + (msg[13] ^ hash[13]);
q[3] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[4];
q[3] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[4];
tmp = (msg[ 1] ^ hash[ 1]) + (msg[ 2] ^ hash[ 2]) + (msg[ 9] ^ hash[ 9]) - (msg[11] ^ hash[11]) - (msg[14] ^ hash[14]);
q[4] = (SHR(tmp, 1) ^ tmp) + hash[5];
tmp = (msg[ 3] ^ hash[ 3]) - (msg[ 2] ^ hash[ 2]) + (msg[10] ^ hash[10]) - (msg[12] ^ hash[12]) + (msg[15] ^ hash[15]);
q[5] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[6];
q[5] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[6];
tmp = (msg[ 4] ^ hash[ 4]) - (msg[ 0] ^ hash[ 0]) - (msg[ 3] ^ hash[ 3]) - (msg[11] ^ hash[11]) + (msg[13] ^ hash[13]);
q[6] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[7];
q[6] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[7];
tmp = (msg[ 1] ^ hash[ 1]) - (msg[ 4] ^ hash[ 4]) - (msg[ 5] ^ hash[ 5]) - (msg[12] ^ hash[12]) - (msg[14] ^ hash[14]);
q[7] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[8];
q[7] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[8];
tmp = (msg[ 2] ^ hash[ 2]) - (msg[ 5] ^ hash[ 5]) - (msg[ 6] ^ hash[ 6]) + (msg[13] ^ hash[13]) - (msg[15] ^ hash[15]);
q[8] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[9];
q[8] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[9];
tmp = (msg[ 0] ^ hash[ 0]) - (msg[ 3] ^ hash[ 3]) + (msg[ 6] ^ hash[ 6]) - (msg[ 7] ^ hash[ 7]) + (msg[14] ^ hash[14]);
q[9] = (SHR(tmp, 1) ^ tmp) + hash[10];
tmp = (msg[ 8] ^ hash[ 8]) - (msg[ 1] ^ hash[ 1]) - (msg[ 4] ^ hash[ 4]) - (msg[ 7] ^ hash[ 7]) + (msg[15] ^ hash[15]);
q[10] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[11];
q[10] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[11];
tmp = (msg[ 8] ^ hash[ 8]) - (msg[ 0] ^ hash[ 0]) - (msg[ 2] ^ hash[ 2]) - (msg[ 5] ^ hash[ 5]) + (msg[ 9] ^ hash[ 9]);
q[11] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROL(tmp, 13) ^ ROL(tmp, 43)) + hash[12];
q[11] = (SHR(tmp, 1) ^ SHL(tmp, 2) ^ ROTL64(tmp, 13) ^ ROTL64(tmp, 43)) + hash[12];
tmp = (msg[ 1] ^ hash[ 1]) + (msg[ 3] ^ hash[ 3]) - (msg[ 6] ^ hash[ 6]) - (msg[ 9] ^ hash[ 9]) + (msg[10] ^ hash[10]);
q[12] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROL(tmp, 19) ^ ROL(tmp, 53)) + hash[13];
q[12] = (SHR(tmp, 2) ^ SHL(tmp, 1) ^ ROTL64(tmp, 19) ^ ROTL64(tmp, 53)) + hash[13];
tmp = (msg[ 2] ^ hash[ 2]) + (msg[ 4] ^ hash[ 4]) + (msg[ 7] ^ hash[ 7]) + (msg[10] ^ hash[10]) + (msg[11] ^ hash[11]);
q[13] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROL(tmp, 28) ^ ROL(tmp, 59)) + hash[14];
q[13] = (SHR(tmp, 2) ^ SHL(tmp, 2) ^ ROTL64(tmp, 28) ^ ROTL64(tmp, 59)) + hash[14];
tmp = (msg[ 3] ^ hash[ 3]) - (msg[ 5] ^ hash[ 5]) + (msg[ 8] ^ hash[ 8]) - (msg[11] ^ hash[11]) - (msg[12] ^ hash[12]);
q[14] = (SHR(tmp, 1) ^ tmp) + hash[15];
tmp = (msg[12] ^ hash[12]) - (msg[ 4] ^ hash[ 4]) - (msg[ 6] ^ hash[ 6]) - (msg[ 9] ^ hash[ 9]) + (msg[13] ^ hash[13]);
q[15] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROL(tmp, 4) ^ ROL(tmp, 37)) + hash[0];
q[15] = (SHR(tmp, 1) ^ SHL(tmp, 3) ^ ROTL64(tmp, 4) ^ ROTL64(tmp, 37)) + hash[0];
q[0+16] =
(SHR(q[0], 1) ^ SHL(q[0], 2) ^ ROL(q[0], 13) ^ ROL(q[0], 43)) +
(SHR(q[0+1], 2) ^ SHL(q[0+1], 1) ^ ROL(q[0+1], 19) ^ ROL(q[0+1], 53)) +
(SHR(q[0+2], 2) ^ SHL(q[0+2], 2) ^ ROL(q[0+2], 28) ^ ROL(q[0+2], 59)) +
(SHR(q[0+3], 1) ^ SHL(q[0+3], 3) ^ ROL(q[0+3], 4) ^ ROL(q[0+3], 37)) +
(SHR(q[0+4], 1) ^ SHL(q[0+4], 2) ^ ROL(q[0+4], 13) ^ ROL(q[0+4], 43)) +
(SHR(q[0+5], 2) ^ SHL(q[0+5], 1) ^ ROL(q[0+5], 19) ^ ROL(q[0+5], 53)) +
(SHR(q[0+6], 2) ^ SHL(q[0+6], 2) ^ ROL(q[0+6], 28) ^ ROL(q[0+6], 59)) +
(SHR(q[0+7], 1) ^ SHL(q[0+7], 3) ^ ROL(q[0+7], 4) ^ ROL(q[0+7], 37)) +
(SHR(q[0+8], 1) ^ SHL(q[0+8], 2) ^ ROL(q[0+8], 13) ^ ROL(q[0+8], 43)) +
(SHR(q[0+9], 2) ^ SHL(q[0+9], 1) ^ ROL(q[0+9], 19) ^ ROL(q[0+9], 53)) +
(SHR(q[0+10], 2) ^ SHL(q[0+10], 2) ^ ROL(q[0+10], 28) ^ ROL(q[0+10], 59)) +
(SHR(q[0+11], 1) ^ SHL(q[0+11], 3) ^ ROL(q[0+11], 4) ^ ROL(q[0+11], 37)) +
(SHR(q[0+12], 1) ^ SHL(q[0+12], 2) ^ ROL(q[0+12], 13) ^ ROL(q[0+12], 43)) +
(SHR(q[0+13], 2) ^ SHL(q[0+13], 1) ^ ROL(q[0+13], 19) ^ ROL(q[0+13], 53)) +
(SHR(q[0+14], 2) ^ SHL(q[0+14], 2) ^ ROL(q[0+14], 28) ^ ROL(q[0+14], 59)) +
(SHR(q[0+15], 1) ^ SHL(q[0+15], 3) ^ ROL(q[0+15], 4) ^ ROL(q[0+15], 37)) +
((make_uint2(0x55555550ul, 0x55555555) + ROL(msg[0], 0 + 1) +
ROL(msg[0+3], 0+4) - ROL(msg[0+10], 0+11) ) ^ hash[0+7]);
(SHR(q[0], 1) ^ SHL(q[0], 2) ^ ROTL64(q[0], 13) ^ ROTL64(q[0], 43)) +
(SHR(q[0+1], 2) ^ SHL(q[0+1], 1) ^ ROTL64(q[0+1], 19) ^ ROTL64(q[0+1], 53)) +
(SHR(q[0+2], 2) ^ SHL(q[0+2], 2) ^ ROTL64(q[0+2], 28) ^ ROTL64(q[0+2], 59)) +
(SHR(q[0+3], 1) ^ SHL(q[0+3], 3) ^ ROTL64(q[0+3], 4) ^ ROTL64(q[0+3], 37)) +
(SHR(q[0+4], 1) ^ SHL(q[0+4], 2) ^ ROTL64(q[0+4], 13) ^ ROTL64(q[0+4], 43)) +
(SHR(q[0+5], 2) ^ SHL(q[0+5], 1) ^ ROTL64(q[0+5], 19) ^ ROTL64(q[0+5], 53)) +
(SHR(q[0+6], 2) ^ SHL(q[0+6], 2) ^ ROTL64(q[0+6], 28) ^ ROTL64(q[0+6], 59)) +
(SHR(q[0+7], 1) ^ SHL(q[0+7], 3) ^ ROTL64(q[0+7], 4) ^ ROTL64(q[0+7], 37)) +
(SHR(q[0+8], 1) ^ SHL(q[0+8], 2) ^ ROTL64(q[0+8], 13) ^ ROTL64(q[0+8], 43)) +
(SHR(q[0+9], 2) ^ SHL(q[0+9], 1) ^ ROTL64(q[0+9], 19) ^ ROTL64(q[0+9], 53)) +
(SHR(q[0+10], 2) ^ SHL(q[0+10], 2) ^ ROTL64(q[0+10], 28) ^ ROTL64(q[0+10], 59)) +
(SHR(q[0+11], 1) ^ SHL(q[0+11], 3) ^ ROTL64(q[0+11], 4) ^ ROTL64(q[0+11], 37)) +
(SHR(q[0+12], 1) ^ SHL(q[0+12], 2) ^ ROTL64(q[0+12], 13) ^ ROTL64(q[0+12], 43)) +
(SHR(q[0+13], 2) ^ SHL(q[0+13], 1) ^ ROTL64(q[0+13], 19) ^ ROTL64(q[0+13], 53)) +
(SHR(q[0+14], 2) ^ SHL(q[0+14], 2) ^ ROTL64(q[0+14], 28) ^ ROTL64(q[0+14], 59)) +
(SHR(q[0+15], 1) ^ SHL(q[0+15], 3) ^ ROTL64(q[0+15], 4) ^ ROTL64(q[0+15], 37)) +
((make_uint2(0x55555550ul, 0x55555555) + ROTL64(msg[0], 0 + 1) +
ROTL64(msg[0+3], 0+4) - ROTL64(msg[0+10], 0+11) ) ^ hash[0+7]);
q[1 + 16] =
(SHR(q[1], 1) ^ SHL(q[1], 2) ^ ROL(q[1], 13) ^ ROL(q[1], 43)) +
(SHR(q[1 + 1], 2) ^ SHL(q[1 + 1], 1) ^ ROL(q[1 + 1], 19) ^ ROL(q[1 + 1], 53)) +
(SHR(q[1 + 2], 2) ^ SHL(q[1 + 2], 2) ^ ROL(q[1 + 2], 28) ^ ROL(q[1 + 2], 59)) +
(SHR(q[1 + 3], 1) ^ SHL(q[1 + 3], 3) ^ ROL(q[1 + 3], 4) ^ ROL(q[1 + 3], 37)) +
(SHR(q[1 + 4], 1) ^ SHL(q[1 + 4], 2) ^ ROL(q[1 + 4], 13) ^ ROL(q[1 + 4], 43)) +
(SHR(q[1 + 5], 2) ^ SHL(q[1 + 5], 1) ^ ROL(q[1 + 5], 19) ^ ROL(q[1 + 5], 53)) +
(SHR(q[1 + 6], 2) ^ SHL(q[1 + 6], 2) ^ ROL(q[1 + 6], 28) ^ ROL(q[1 + 6], 59)) +
(SHR(q[1 + 7], 1) ^ SHL(q[1 + 7], 3) ^ ROL(q[1 + 7], 4) ^ ROL(q[1 + 7], 37)) +
(SHR(q[1 + 8], 1) ^ SHL(q[1 + 8], 2) ^ ROL(q[1 + 8], 13) ^ ROL(q[1 + 8], 43)) +
(SHR(q[1 + 9], 2) ^ SHL(q[1 + 9], 1) ^ ROL(q[1 + 9], 19) ^ ROL(q[1 + 9], 53)) +
(SHR(q[1 + 10], 2) ^ SHL(q[1 + 10], 2) ^ ROL(q[1 + 10], 28) ^ ROL(q[1 + 10], 59)) +
(SHR(q[1 + 11], 1) ^ SHL(q[1 + 11], 3) ^ ROL(q[1 + 11], 4) ^ ROL(q[1 + 11], 37)) +
(SHR(q[1 + 12], 1) ^ SHL(q[1 + 12], 2) ^ ROL(q[1 + 12], 13) ^ ROL(q[1 + 12], 43)) +
(SHR(q[1 + 13], 2) ^ SHL(q[1 + 13], 1) ^ ROL(q[1 + 13], 19) ^ ROL(q[1 + 13], 53)) +
(SHR(q[1 + 14], 2) ^ SHL(q[1 + 14], 2) ^ ROL(q[1 + 14], 28) ^ ROL(q[1 + 14], 59)) +
(SHR(q[1 + 15], 1) ^ SHL(q[1 + 15], 3) ^ ROL(q[1 + 15], 4) ^ ROL(q[1 + 15], 37)) +
((make_uint2(0xAAAAAAA5, 0x5AAAAAAA) + ROL(msg[1], 1 + 1) +
ROL(msg[1 + 3], 1 + 4) - ROL(msg[1 + 10], 1 + 11)) ^ hash[1 + 7]);
(SHR(q[1], 1) ^ SHL(q[1], 2) ^ ROTL64(q[1], 13) ^ ROTL64(q[1], 43)) +
(SHR(q[1 + 1], 2) ^ SHL(q[1 + 1], 1) ^ ROTL64(q[1 + 1], 19) ^ ROTL64(q[1 + 1], 53)) +
(SHR(q[1 + 2], 2) ^ SHL(q[1 + 2], 2) ^ ROTL64(q[1 + 2], 28) ^ ROTL64(q[1 + 2], 59)) +
(SHR(q[1 + 3], 1) ^ SHL(q[1 + 3], 3) ^ ROTL64(q[1 + 3], 4) ^ ROTL64(q[1 + 3], 37)) +
(SHR(q[1 + 4], 1) ^ SHL(q[1 + 4], 2) ^ ROTL64(q[1 + 4], 13) ^ ROTL64(q[1 + 4], 43)) +
(SHR(q[1 + 5], 2) ^ SHL(q[1 + 5], 1) ^ ROTL64(q[1 + 5], 19) ^ ROTL64(q[1 + 5], 53)) +
(SHR(q[1 + 6], 2) ^ SHL(q[1 + 6], 2) ^ ROTL64(q[1 + 6], 28) ^ ROTL64(q[1 + 6], 59)) +
(SHR(q[1 + 7], 1) ^ SHL(q[1 + 7], 3) ^ ROTL64(q[1 + 7], 4) ^ ROTL64(q[1 + 7], 37)) +
(SHR(q[1 + 8], 1) ^ SHL(q[1 + 8], 2) ^ ROTL64(q[1 + 8], 13) ^ ROTL64(q[1 + 8], 43)) +
(SHR(q[1 + 9], 2) ^ SHL(q[1 + 9], 1) ^ ROTL64(q[1 + 9], 19) ^ ROTL64(q[1 + 9], 53)) +
(SHR(q[1 + 10], 2) ^ SHL(q[1 + 10], 2) ^ ROTL64(q[1 + 10], 28) ^ ROTL64(q[1 + 10], 59)) +
(SHR(q[1 + 11], 1) ^ SHL(q[1 + 11], 3) ^ ROTL64(q[1 + 11], 4) ^ ROTL64(q[1 + 11], 37)) +
(SHR(q[1 + 12], 1) ^ SHL(q[1 + 12], 2) ^ ROTL64(q[1 + 12], 13) ^ ROTL64(q[1 + 12], 43)) +
(SHR(q[1 + 13], 2) ^ SHL(q[1 + 13], 1) ^ ROTL64(q[1 + 13], 19) ^ ROTL64(q[1 + 13], 53)) +
(SHR(q[1 + 14], 2) ^ SHL(q[1 + 14], 2) ^ ROTL64(q[1 + 14], 28) ^ ROTL64(q[1 + 14], 59)) +
(SHR(q[1 + 15], 1) ^ SHL(q[1 + 15], 3) ^ ROTL64(q[1 + 15], 4) ^ ROTL64(q[1 + 15], 37)) +
((make_uint2(0xAAAAAAA5, 0x5AAAAAAA) + ROTL64(msg[1], 1 + 1) +
ROTL64(msg[1 + 3], 1 + 4) - ROTL64(msg[1 + 10], 1 + 11)) ^ hash[1 + 7]);
q[2 + 16] = CONST_EXP2(2) +
((make_uint2(0xFFFFFFFA, 0x5FFFFFFF) + ROL(msg[2], 2 + 1) +
ROL(msg[2+3], 2+4) - ROL(msg[2+10], 2+11) ) ^ hash[2+7]);
((make_uint2(0xFFFFFFFA, 0x5FFFFFFF) + ROTL64(msg[2], 2 + 1) +
ROTL64(msg[2+3], 2+4) - ROTL64(msg[2+10], 2+11) ) ^ hash[2+7]);
q[3 + 16] = CONST_EXP2(3) +
((make_uint2(0x5555554F, 0x65555555) + ROL(msg[3], 3 + 1) +
ROL(msg[3 + 3], 3 + 4) - ROL(msg[3 + 10], 3 + 11)) ^ hash[3 + 7]);
((make_uint2(0x5555554F, 0x65555555) + ROTL64(msg[3], 3 + 1) +
ROTL64(msg[3 + 3], 3 + 4) - ROTL64(msg[3 + 10], 3 + 11)) ^ hash[3 + 7]);
q[4 + 16] = CONST_EXP2(4) +
((make_uint2(0xAAAAAAA4, 0x6AAAAAAA) + ROL(msg[4], 4 + 1) +
ROL(msg[4 + 3], 4 + 4) - ROL(msg[4 + 10], 4 + 11)) ^ hash[4 + 7]);
((make_uint2(0xAAAAAAA4, 0x6AAAAAAA) + ROTL64(msg[4], 4 + 1) +
ROTL64(msg[4 + 3], 4 + 4) - ROTL64(msg[4 + 10], 4 + 11)) ^ hash[4 + 7]);
q[5 + 16] = CONST_EXP2(5) +
((make_uint2(0xFFFFFFF9, 0x6FFFFFFF) + ROL(msg[5], 5 + 1) +
ROL(msg[5 + 3], 5 + 4) - ROL(msg[5 + 10], 5 + 11)) ^ hash[5 + 7]);
((make_uint2(0xFFFFFFF9, 0x6FFFFFFF) + ROTL64(msg[5], 5 + 1) +
ROTL64(msg[5 + 3], 5 + 4) - ROTL64(msg[5 + 10], 5 + 11)) ^ hash[5 + 7]);
q[6 + 16] = CONST_EXP2(6) +
((make_uint2(0x5555554E, 0x75555555)+ ROL(msg[6], 6 + 1) +
ROL(msg[6 + 3], 6 + 4) - ROL(msg[6 - 6], (6 - 6) + 1)) ^ hash[6 + 7]);
((make_uint2(0x5555554E, 0x75555555)+ ROTL64(msg[6], 6 + 1) +
ROTL64(msg[6 + 3], 6 + 4) - ROTL64(msg[6 - 6], (6 - 6) + 1)) ^ hash[6 + 7]);
q[7 + 16] = CONST_EXP2(7) +
((make_uint2(0xAAAAAAA3, 0x7AAAAAAA) + ROL(msg[7], 7 + 1) +
ROL(msg[7 + 3], 7 + 4) - ROL(msg[7 - 6], (7 - 6) + 1)) ^ hash[7 + 7]);
((make_uint2(0xAAAAAAA3, 0x7AAAAAAA) + ROTL64(msg[7], 7 + 1) +
ROTL64(msg[7 + 3], 7 + 4) - ROTL64(msg[7 - 6], (7 - 6) + 1)) ^ hash[7 + 7]);
q[8 + 16] = CONST_EXP2(8) +
((make_uint2(0xFFFFFFF8, 0x7FFFFFFF) + ROL(msg[8], 8 + 1) +
ROL(msg[8 + 3], 8 + 4) - ROL(msg[8 - 6], (8 - 6) + 1)) ^ hash[8 + 7]);
((make_uint2(0xFFFFFFF8, 0x7FFFFFFF) + ROTL64(msg[8], 8 + 1) +
ROTL64(msg[8 + 3], 8 + 4) - ROTL64(msg[8 - 6], (8 - 6) + 1)) ^ hash[8 + 7]);
q[9 + 16] = CONST_EXP2(9) +
((make_uint2(0x5555554D, 0x85555555) + ROL(msg[9], 9 + 1) +
ROL(msg[9 + 3], 9 + 4) - ROL(msg[9 - 6], (9 - 6) + 1)) ^ hash[9 - 9]);
((make_uint2(0x5555554D, 0x85555555) + ROTL64(msg[9], 9 + 1) +
ROTL64(msg[9 + 3], 9 + 4) - ROTL64(msg[9 - 6], (9 - 6) + 1)) ^ hash[9 - 9]);
q[10 + 16] = CONST_EXP2(10) +
((make_uint2(0xAAAAAAA2, 0x8AAAAAAA) + ROL(msg[10], 10 + 1) +
ROL(msg[10 + 3], 10 + 4) - ROL(msg[10 - 6], (10 - 6) + 1)) ^ hash[10 - 9]);
((make_uint2(0xAAAAAAA2, 0x8AAAAAAA) + ROTL64(msg[10], 10 + 1) +
ROTL64(msg[10 + 3], 10 + 4) - ROTL64(msg[10 - 6], (10 - 6) + 1)) ^ hash[10 - 9]);
q[11 + 16] = CONST_EXP2(11) +
((make_uint2(0xFFFFFFF7, 0x8FFFFFFF) + ROL(msg[11], 11 + 1) +
ROL(msg[11 + 3], 11 + 4) - ROL(msg[11 - 6], (11 - 6) + 1)) ^ hash[11 - 9]);
((make_uint2(0xFFFFFFF7, 0x8FFFFFFF) + ROTL64(msg[11], 11 + 1) +
ROTL64(msg[11 + 3], 11 + 4) - ROTL64(msg[11 - 6], (11 - 6) + 1)) ^ hash[11 - 9]);
q[12 + 16] = CONST_EXP2(12) +
((make_uint2(0x5555554C, 0x95555555) + ROL(msg[12], 12 + 1) +
ROL(msg[12 + 3], 12 + 4) - ROL(msg[12 - 6], (12 - 6) + 1)) ^ hash[12 - 9]);
((make_uint2(0x5555554C, 0x95555555) + ROTL64(msg[12], 12 + 1) +
ROTL64(msg[12 + 3], 12 + 4) - ROTL64(msg[12 - 6], (12 - 6) + 1)) ^ hash[12 - 9]);
q[13 + 16] = CONST_EXP2(13) +
((make_uint2(0xAAAAAAA1, 0x9AAAAAAA) + ROL(msg[13], 13 + 1) +
ROL(msg[13 - 13], (13 - 13) + 1) - ROL(msg[13 - 6], (13 - 6) + 1)) ^ hash[13 - 9]);
((make_uint2(0xAAAAAAA1, 0x9AAAAAAA) + ROTL64(msg[13], 13 + 1) +
ROTL64(msg[13 - 13], (13 - 13) + 1) - ROTL64(msg[13 - 6], (13 - 6) + 1)) ^ hash[13 - 9]);
q[14 + 16] = CONST_EXP2(14) +
((make_uint2(0xFFFFFFF6, 0x9FFFFFFF) + ROL(msg[14], 14 + 1) +
ROL(msg[14 - 13], (14 - 13) + 1) - ROL(msg[14 - 6], (14 - 6) + 1)) ^ hash[14 - 9]);
((make_uint2(0xFFFFFFF6, 0x9FFFFFFF) + ROTL64(msg[14], 14 + 1) +
ROTL64(msg[14 - 13], (14 - 13) + 1) - ROTL64(msg[14 - 6], (14 - 6) + 1)) ^ hash[14 - 9]);
q[15 + 16] = CONST_EXP2(15) +
((make_uint2(0x5555554B, 0xA5555555) + ROL(msg[15], 15 + 1) +
ROL(msg[15 - 13], (15 - 13) + 1) - ROL(msg[15 - 6], (15 - 6) + 1)) ^ hash[15 - 9]);
((make_uint2(0x5555554B, 0xA5555555) + ROTL64(msg[15], 15 + 1) +
ROTL64(msg[15 - 13], (15 - 13) + 1) - ROTL64(msg[15 - 6], (15 - 6) + 1)) ^ hash[15 - 9]);
uint2 XL64 = q[16]^q[17]^q[18]^q[19]^q[20]^q[21]^q[22]^q[23];
uint2 XH64 = XL64^q[24] ^ q[25] ^ q[26] ^ q[27] ^ q[28] ^ q[29] ^ q[30] ^ q[31];
hash[0] = (SHL(XH64, 5) ^ SHR(q[16],5) ^ msg[ 0]) + (XL64 ^ q[24] ^ q[ 0]);
hash[1] = (SHR(XH64, 7) ^ SHL(q[17],8) ^ msg[ 1]) + (XL64 ^ q[25] ^ q[ 1]);
hash[2] = (SHR(XH64, 5) ^ SHL(q[18],5) ^ msg[ 2]) + (XL64 ^ q[26] ^ q[ 2]);
hash[3] = (SHR(XH64, 1) ^ SHL(q[19],5) ^ msg[ 3]) + (XL64 ^ q[27] ^ q[ 3]);
hash[4] = (SHR(XH64, 3) ^ q[20] ^ msg[ 4]) + (XL64 ^ q[28] ^ q[ 4]);
hash[5] = (SHL(XH64, 6) ^ SHR(q[21],6) ^ msg[ 5]) + (XL64 ^ q[29] ^ q[ 5]);
hash[6] = (SHR(XH64, 4) ^ SHL(q[22],6) ^ msg[ 6]) + (XL64 ^ q[30] ^ q[ 6]);
hash[7] = (SHR(XH64,11) ^ SHL(q[23],2) ^ msg[ 7]) + (XL64 ^ q[31] ^ q[ 7]);
hash[0] = (SHL(XH64, 5) ^ SHR(q[16],5) ^ msg[ 0]) + ( XL64 ^ q[24] ^ q[ 0]);
hash[1] = (SHR(XH64, 7) ^ SHL(q[17],8) ^ msg[ 1]) + ( XL64 ^ q[25] ^ q[ 1]);
hash[2] = (SHR(XH64, 5) ^ SHL(q[18],5) ^ msg[ 2]) + ( XL64 ^ q[26] ^ q[ 2]);
hash[3] = (SHR(XH64, 1) ^ SHL(q[19],5) ^ msg[ 3]) + ( XL64 ^ q[27] ^ q[ 3]);
hash[4] = (SHR(XH64, 3) ^ q[20] ^ msg[ 4]) + ( XL64 ^ q[28] ^ q[ 4]);
hash[5] = (SHL(XH64, 6) ^ SHR(q[21],6) ^ msg[ 5]) + ( XL64 ^ q[29] ^ q[ 5]);
hash[6] = (SHR(XH64, 4) ^ SHL(q[22],6) ^ msg[ 6]) + ( XL64 ^ q[30] ^ q[ 6]);
hash[7] = (SHR(XH64,11) ^ SHL(q[23],2) ^ msg[ 7]) + ( XL64 ^ q[31] ^ q[ 7]);
hash[ 8] = ROL(hash[4], 9) + (XH64 ^ q[24] ^ msg[ 8]) + (SHL(XL64,8) ^ q[23] ^ q[ 8]);
hash[ 9] = ROL(hash[5],10) + (XH64 ^ q[25] ^ msg[ 9]) + (SHR(XL64,6) ^ q[16] ^ q[ 9]);
hash[10] = ROL(hash[6],11) + (XH64 ^ q[26] ^ msg[10]) + (SHL(XL64,6) ^ q[17] ^ q[10]);
hash[11] = ROL(hash[7],12) + (XH64 ^ q[27] ^ msg[11]) + (SHL(XL64,4) ^ q[18] ^ q[11]);
hash[12] = ROL(hash[0],13) + (XH64 ^ q[28] ^ msg[12]) + (SHR(XL64,3) ^ q[19] ^ q[12]);
hash[13] = ROL(hash[1],14) + (XH64 ^ q[29] ^ msg[13]) + (SHR(XL64,4) ^ q[20] ^ q[13]);
hash[14] = ROL(hash[2],15) + (XH64 ^ q[30] ^ msg[14]) + (SHR(XL64,7) ^ q[21] ^ q[14]);
hash[15] = ROL(hash[3],16) + (XH64 ^ q[31] ^ msg[15]) + (SHR(XL64, 2) ^ q[22] ^ q[15]);
hash[ 8] = ROTL64(hash[4], 9) + ( XH64 ^ q[24] ^ msg[ 8]) + (SHL(XL64,8) ^ q[23] ^ q[ 8]);
hash[ 9] = ROTL64(hash[5],10) + ( XH64 ^ q[25] ^ msg[ 9]) + (SHR(XL64,6) ^ q[16] ^ q[ 9]);
hash[10] = ROTL64(hash[6],11) + ( XH64 ^ q[26] ^ msg[10]) + (SHL(XL64,6) ^ q[17] ^ q[10]);
hash[11] = ROTL64(hash[7],12) + ( XH64 ^ q[27] ^ msg[11]) + (SHL(XL64,4) ^ q[18] ^ q[11]);
hash[12] = ROTL64(hash[0],13) + ( XH64 ^ q[28] ^ msg[12]) + (SHR(XL64,3) ^ q[19] ^ q[12]);
hash[13] = ROTL64(hash[1],14) + ( XH64 ^ q[29] ^ msg[13]) + (SHR(XL64,4) ^ q[20] ^ q[13]);
hash[14] = ROTL64(hash[2],15) + ( XH64 ^ q[30] ^ msg[14]) + (SHR(XL64,7) ^ q[21] ^ q[14]);
hash[15] = ROTL64(hash[3],16) + (XH64 ^ q[31] ^ msg[15]) + (SHR(XL64, 2) ^ q[22] ^ q[15]);
}
__global__
@@ -320,8 +323,8 @@ void quark_bmw512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t *
{
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
uint32_t hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[hashPosition * 8];
int hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[8 * hashPosition];
// Init
uint2 h[16] = {
@@ -346,11 +349,10 @@ void quark_bmw512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t *
// Nachricht kopieren (Achtung, die Nachricht hat 64 Byte,
// BMW arbeitet mit 128 Byte!!!
uint2 message[16];
#pragma unroll
#pragma unroll 8
for(int i=0;i<8;i++)
message[i] = vectorize(inpHash[i]);
#pragma unroll 6
#pragma unroll 6
for(int i=9;i<15;i++)
message[i] = make_uint2(0,0);
@@ -363,7 +365,7 @@ void quark_bmw512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t *
Compression512_64_first(message, h);
// Final
#pragma unroll
#pragma unroll 16
for(int i=0;i<16;i++)
{
message[i].y = 0xaaaaaaaa;
@@ -372,9 +374,9 @@ void quark_bmw512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t *
Compression512(h, message);
// fertig
uint64_t *outpHash = &g_hash[hashPosition * 8];
uint64_t *outpHash = &g_hash[8 * hashPosition];
#pragma unroll
#pragma unroll 8
for(int i=0;i<8;i++)
outpHash[i] = devectorize(message[i+8]);
}
@@ -428,7 +430,7 @@ void quark_bmw512_gpu_hash_80(uint32_t threads, uint32_t startNounce, uint64_t *
Compression512(h, message);
// fertig
uint64_t *outpHash = &g_hash[thread * 8];
uint64_t *outpHash = &g_hash[8 * thread];
#pragma unroll 8
for(int i=0;i<8;i++)
+3 -3
View File
@@ -151,8 +151,8 @@ void quark_bmw512_gpu_hash_64_30(uint32_t threads, uint32_t startNounce, uint64_
{
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
uint32_t hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[hashPosition * 8];
int hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[8 * hashPosition];
// Init
uint64_t h[16];
@@ -205,7 +205,7 @@ void quark_bmw512_gpu_hash_64_30(uint32_t threads, uint32_t startNounce, uint64_
Compression512_30(h, message);
uint64_t *outpHash = &g_hash[hashPosition * 8];
uint64_t *outpHash = &g_hash[8 * hashPosition];
#pragma unroll 8
for(int i=0;i<8;i++)
+204 -247
View File
@@ -1,287 +1,245 @@
#include "cuda_helper.h"
// #include <stdio.h> // printf
// #include <unistd.h> // sleep
__constant__ unsigned char c_E8_bitslice_roundconstant[42][32] = {
{ 0x72, 0xd5, 0xde, 0xa2, 0xdf, 0x15, 0xf8, 0x67, 0x7b, 0x84, 0x15, 0xa, 0xb7, 0x23, 0x15, 0x57, 0x81, 0xab, 0xd6, 0x90, 0x4d, 0x5a, 0x87, 0xf6, 0x4e, 0x9f, 0x4f, 0xc5, 0xc3, 0xd1, 0x2b, 0x40 },
{ 0xea, 0x98, 0x3a, 0xe0, 0x5c, 0x45, 0xfa, 0x9c, 0x3, 0xc5, 0xd2, 0x99, 0x66, 0xb2, 0x99, 0x9a, 0x66, 0x2, 0x96, 0xb4, 0xf2, 0xbb, 0x53, 0x8a, 0xb5, 0x56, 0x14, 0x1a, 0x88, 0xdb, 0xa2, 0x31 },
{ 0x3, 0xa3, 0x5a, 0x5c, 0x9a, 0x19, 0xe, 0xdb, 0x40, 0x3f, 0xb2, 0xa, 0x87, 0xc1, 0x44, 0x10, 0x1c, 0x5, 0x19, 0x80, 0x84, 0x9e, 0x95, 0x1d, 0x6f, 0x33, 0xeb, 0xad, 0x5e, 0xe7, 0xcd, 0xdc },
{ 0x10, 0xba, 0x13, 0x92, 0x2, 0xbf, 0x6b, 0x41, 0xdc, 0x78, 0x65, 0x15, 0xf7, 0xbb, 0x27, 0xd0, 0xa, 0x2c, 0x81, 0x39, 0x37, 0xaa, 0x78, 0x50, 0x3f, 0x1a, 0xbf, 0xd2, 0x41, 0x0, 0x91, 0xd3 },
{ 0x42, 0x2d, 0x5a, 0xd, 0xf6, 0xcc, 0x7e, 0x90, 0xdd, 0x62, 0x9f, 0x9c, 0x92, 0xc0, 0x97, 0xce, 0x18, 0x5c, 0xa7, 0xb, 0xc7, 0x2b, 0x44, 0xac, 0xd1, 0xdf, 0x65, 0xd6, 0x63, 0xc6, 0xfc, 0x23 },
{ 0x97, 0x6e, 0x6c, 0x3, 0x9e, 0xe0, 0xb8, 0x1a, 0x21, 0x5, 0x45, 0x7e, 0x44, 0x6c, 0xec, 0xa8, 0xee, 0xf1, 0x3, 0xbb, 0x5d, 0x8e, 0x61, 0xfa, 0xfd, 0x96, 0x97, 0xb2, 0x94, 0x83, 0x81, 0x97 },
{ 0x4a, 0x8e, 0x85, 0x37, 0xdb, 0x3, 0x30, 0x2f, 0x2a, 0x67, 0x8d, 0x2d, 0xfb, 0x9f, 0x6a, 0x95, 0x8a, 0xfe, 0x73, 0x81, 0xf8, 0xb8, 0x69, 0x6c, 0x8a, 0xc7, 0x72, 0x46, 0xc0, 0x7f, 0x42, 0x14 },
{ 0xc5, 0xf4, 0x15, 0x8f, 0xbd, 0xc7, 0x5e, 0xc4, 0x75, 0x44, 0x6f, 0xa7, 0x8f, 0x11, 0xbb, 0x80, 0x52, 0xde, 0x75, 0xb7, 0xae, 0xe4, 0x88, 0xbc, 0x82, 0xb8, 0x0, 0x1e, 0x98, 0xa6, 0xa3, 0xf4 },
{ 0x8e, 0xf4, 0x8f, 0x33, 0xa9, 0xa3, 0x63, 0x15, 0xaa, 0x5f, 0x56, 0x24, 0xd5, 0xb7, 0xf9, 0x89, 0xb6, 0xf1, 0xed, 0x20, 0x7c, 0x5a, 0xe0, 0xfd, 0x36, 0xca, 0xe9, 0x5a, 0x6, 0x42, 0x2c, 0x36 },
{ 0xce, 0x29, 0x35, 0x43, 0x4e, 0xfe, 0x98, 0x3d, 0x53, 0x3a, 0xf9, 0x74, 0x73, 0x9a, 0x4b, 0xa7, 0xd0, 0xf5, 0x1f, 0x59, 0x6f, 0x4e, 0x81, 0x86, 0xe, 0x9d, 0xad, 0x81, 0xaf, 0xd8, 0x5a, 0x9f },
{ 0xa7, 0x5, 0x6, 0x67, 0xee, 0x34, 0x62, 0x6a, 0x8b, 0xb, 0x28, 0xbe, 0x6e, 0xb9, 0x17, 0x27, 0x47, 0x74, 0x7, 0x26, 0xc6, 0x80, 0x10, 0x3f, 0xe0, 0xa0, 0x7e, 0x6f, 0xc6, 0x7e, 0x48, 0x7b },
{ 0xd, 0x55, 0xa, 0xa5, 0x4a, 0xf8, 0xa4, 0xc0, 0x91, 0xe3, 0xe7, 0x9f, 0x97, 0x8e, 0xf1, 0x9e, 0x86, 0x76, 0x72, 0x81, 0x50, 0x60, 0x8d, 0xd4, 0x7e, 0x9e, 0x5a, 0x41, 0xf3, 0xe5, 0xb0, 0x62 },
{ 0xfc, 0x9f, 0x1f, 0xec, 0x40, 0x54, 0x20, 0x7a, 0xe3, 0xe4, 0x1a, 0x0, 0xce, 0xf4, 0xc9, 0x84, 0x4f, 0xd7, 0x94, 0xf5, 0x9d, 0xfa, 0x95, 0xd8, 0x55, 0x2e, 0x7e, 0x11, 0x24, 0xc3, 0x54, 0xa5 },
{ 0x5b, 0xdf, 0x72, 0x28, 0xbd, 0xfe, 0x6e, 0x28, 0x78, 0xf5, 0x7f, 0xe2, 0xf, 0xa5, 0xc4, 0xb2, 0x5, 0x89, 0x7c, 0xef, 0xee, 0x49, 0xd3, 0x2e, 0x44, 0x7e, 0x93, 0x85, 0xeb, 0x28, 0x59, 0x7f },
{ 0x70, 0x5f, 0x69, 0x37, 0xb3, 0x24, 0x31, 0x4a, 0x5e, 0x86, 0x28, 0xf1, 0x1d, 0xd6, 0xe4, 0x65, 0xc7, 0x1b, 0x77, 0x4, 0x51, 0xb9, 0x20, 0xe7, 0x74, 0xfe, 0x43, 0xe8, 0x23, 0xd4, 0x87, 0x8a },
{ 0x7d, 0x29, 0xe8, 0xa3, 0x92, 0x76, 0x94, 0xf2, 0xdd, 0xcb, 0x7a, 0x9, 0x9b, 0x30, 0xd9, 0xc1, 0x1d, 0x1b, 0x30, 0xfb, 0x5b, 0xdc, 0x1b, 0xe0, 0xda, 0x24, 0x49, 0x4f, 0xf2, 0x9c, 0x82, 0xbf },
{ 0xa4, 0xe7, 0xba, 0x31, 0xb4, 0x70, 0xbf, 0xff, 0xd, 0x32, 0x44, 0x5, 0xde, 0xf8, 0xbc, 0x48, 0x3b, 0xae, 0xfc, 0x32, 0x53, 0xbb, 0xd3, 0x39, 0x45, 0x9f, 0xc3, 0xc1, 0xe0, 0x29, 0x8b, 0xa0 },
{ 0xe5, 0xc9, 0x5, 0xfd, 0xf7, 0xae, 0x9, 0xf, 0x94, 0x70, 0x34, 0x12, 0x42, 0x90, 0xf1, 0x34, 0xa2, 0x71, 0xb7, 0x1, 0xe3, 0x44, 0xed, 0x95, 0xe9, 0x3b, 0x8e, 0x36, 0x4f, 0x2f, 0x98, 0x4a },
{ 0x88, 0x40, 0x1d, 0x63, 0xa0, 0x6c, 0xf6, 0x15, 0x47, 0xc1, 0x44, 0x4b, 0x87, 0x52, 0xaf, 0xff, 0x7e, 0xbb, 0x4a, 0xf1, 0xe2, 0xa, 0xc6, 0x30, 0x46, 0x70, 0xb6, 0xc5, 0xcc, 0x6e, 0x8c, 0xe6 },
{ 0xa4, 0xd5, 0xa4, 0x56, 0xbd, 0x4f, 0xca, 0x0, 0xda, 0x9d, 0x84, 0x4b, 0xc8, 0x3e, 0x18, 0xae, 0x73, 0x57, 0xce, 0x45, 0x30, 0x64, 0xd1, 0xad, 0xe8, 0xa6, 0xce, 0x68, 0x14, 0x5c, 0x25, 0x67 },
{ 0xa3, 0xda, 0x8c, 0xf2, 0xcb, 0xe, 0xe1, 0x16, 0x33, 0xe9, 0x6, 0x58, 0x9a, 0x94, 0x99, 0x9a, 0x1f, 0x60, 0xb2, 0x20, 0xc2, 0x6f, 0x84, 0x7b, 0xd1, 0xce, 0xac, 0x7f, 0xa0, 0xd1, 0x85, 0x18 },
{ 0x32, 0x59, 0x5b, 0xa1, 0x8d, 0xdd, 0x19, 0xd3, 0x50, 0x9a, 0x1c, 0xc0, 0xaa, 0xa5, 0xb4, 0x46, 0x9f, 0x3d, 0x63, 0x67, 0xe4, 0x4, 0x6b, 0xba, 0xf6, 0xca, 0x19, 0xab, 0xb, 0x56, 0xee, 0x7e },
{ 0x1f, 0xb1, 0x79, 0xea, 0xa9, 0x28, 0x21, 0x74, 0xe9, 0xbd, 0xf7, 0x35, 0x3b, 0x36, 0x51, 0xee, 0x1d, 0x57, 0xac, 0x5a, 0x75, 0x50, 0xd3, 0x76, 0x3a, 0x46, 0xc2, 0xfe, 0xa3, 0x7d, 0x70, 0x1 },
{ 0xf7, 0x35, 0xc1, 0xaf, 0x98, 0xa4, 0xd8, 0x42, 0x78, 0xed, 0xec, 0x20, 0x9e, 0x6b, 0x67, 0x79, 0x41, 0x83, 0x63, 0x15, 0xea, 0x3a, 0xdb, 0xa8, 0xfa, 0xc3, 0x3b, 0x4d, 0x32, 0x83, 0x2c, 0x83 },
{ 0xa7, 0x40, 0x3b, 0x1f, 0x1c, 0x27, 0x47, 0xf3, 0x59, 0x40, 0xf0, 0x34, 0xb7, 0x2d, 0x76, 0x9a, 0xe7, 0x3e, 0x4e, 0x6c, 0xd2, 0x21, 0x4f, 0xfd, 0xb8, 0xfd, 0x8d, 0x39, 0xdc, 0x57, 0x59, 0xef },
{ 0x8d, 0x9b, 0xc, 0x49, 0x2b, 0x49, 0xeb, 0xda, 0x5b, 0xa2, 0xd7, 0x49, 0x68, 0xf3, 0x70, 0xd, 0x7d, 0x3b, 0xae, 0xd0, 0x7a, 0x8d, 0x55, 0x84, 0xf5, 0xa5, 0xe9, 0xf0, 0xe4, 0xf8, 0x8e, 0x65 },
{ 0xa0, 0xb8, 0xa2, 0xf4, 0x36, 0x10, 0x3b, 0x53, 0xc, 0xa8, 0x7, 0x9e, 0x75, 0x3e, 0xec, 0x5a, 0x91, 0x68, 0x94, 0x92, 0x56, 0xe8, 0x88, 0x4f, 0x5b, 0xb0, 0x5c, 0x55, 0xf8, 0xba, 0xbc, 0x4c },
{ 0xe3, 0xbb, 0x3b, 0x99, 0xf3, 0x87, 0x94, 0x7b, 0x75, 0xda, 0xf4, 0xd6, 0x72, 0x6b, 0x1c, 0x5d, 0x64, 0xae, 0xac, 0x28, 0xdc, 0x34, 0xb3, 0x6d, 0x6c, 0x34, 0xa5, 0x50, 0xb8, 0x28, 0xdb, 0x71 },
{ 0xf8, 0x61, 0xe2, 0xf2, 0x10, 0x8d, 0x51, 0x2a, 0xe3, 0xdb, 0x64, 0x33, 0x59, 0xdd, 0x75, 0xfc, 0x1c, 0xac, 0xbc, 0xf1, 0x43, 0xce, 0x3f, 0xa2, 0x67, 0xbb, 0xd1, 0x3c, 0x2, 0xe8, 0x43, 0xb0 },
{ 0x33, 0xa, 0x5b, 0xca, 0x88, 0x29, 0xa1, 0x75, 0x7f, 0x34, 0x19, 0x4d, 0xb4, 0x16, 0x53, 0x5c, 0x92, 0x3b, 0x94, 0xc3, 0xe, 0x79, 0x4d, 0x1e, 0x79, 0x74, 0x75, 0xd7, 0xb6, 0xee, 0xaf, 0x3f },
{ 0xea, 0xa8, 0xd4, 0xf7, 0xbe, 0x1a, 0x39, 0x21, 0x5c, 0xf4, 0x7e, 0x9, 0x4c, 0x23, 0x27, 0x51, 0x26, 0xa3, 0x24, 0x53, 0xba, 0x32, 0x3c, 0xd2, 0x44, 0xa3, 0x17, 0x4a, 0x6d, 0xa6, 0xd5, 0xad },
{ 0xb5, 0x1d, 0x3e, 0xa6, 0xaf, 0xf2, 0xc9, 0x8, 0x83, 0x59, 0x3d, 0x98, 0x91, 0x6b, 0x3c, 0x56, 0x4c, 0xf8, 0x7c, 0xa1, 0x72, 0x86, 0x60, 0x4d, 0x46, 0xe2, 0x3e, 0xcc, 0x8, 0x6e, 0xc7, 0xf6 },
{ 0x2f, 0x98, 0x33, 0xb3, 0xb1, 0xbc, 0x76, 0x5e, 0x2b, 0xd6, 0x66, 0xa5, 0xef, 0xc4, 0xe6, 0x2a, 0x6, 0xf4, 0xb6, 0xe8, 0xbe, 0xc1, 0xd4, 0x36, 0x74, 0xee, 0x82, 0x15, 0xbc, 0xef, 0x21, 0x63 },
{ 0xfd, 0xc1, 0x4e, 0xd, 0xf4, 0x53, 0xc9, 0x69, 0xa7, 0x7d, 0x5a, 0xc4, 0x6, 0x58, 0x58, 0x26, 0x7e, 0xc1, 0x14, 0x16, 0x6, 0xe0, 0xfa, 0x16, 0x7e, 0x90, 0xaf, 0x3d, 0x28, 0x63, 0x9d, 0x3f },
{ 0xd2, 0xc9, 0xf2, 0xe3, 0x0, 0x9b, 0xd2, 0xc, 0x5f, 0xaa, 0xce, 0x30, 0xb7, 0xd4, 0xc, 0x30, 0x74, 0x2a, 0x51, 0x16, 0xf2, 0xe0, 0x32, 0x98, 0xd, 0xeb, 0x30, 0xd8, 0xe3, 0xce, 0xf8, 0x9a },
{ 0x4b, 0xc5, 0x9e, 0x7b, 0xb5, 0xf1, 0x79, 0x92, 0xff, 0x51, 0xe6, 0x6e, 0x4, 0x86, 0x68, 0xd3, 0x9b, 0x23, 0x4d, 0x57, 0xe6, 0x96, 0x67, 0x31, 0xcc, 0xe6, 0xa6, 0xf3, 0x17, 0xa, 0x75, 0x5 },
{ 0xb1, 0x76, 0x81, 0xd9, 0x13, 0x32, 0x6c, 0xce, 0x3c, 0x17, 0x52, 0x84, 0xf8, 0x5, 0xa2, 0x62, 0xf4, 0x2b, 0xcb, 0xb3, 0x78, 0x47, 0x15, 0x47, 0xff, 0x46, 0x54, 0x82, 0x23, 0x93, 0x6a, 0x48 },
{ 0x38, 0xdf, 0x58, 0x7, 0x4e, 0x5e, 0x65, 0x65, 0xf2, 0xfc, 0x7c, 0x89, 0xfc, 0x86, 0x50, 0x8e, 0x31, 0x70, 0x2e, 0x44, 0xd0, 0xb, 0xca, 0x86, 0xf0, 0x40, 0x9, 0xa2, 0x30, 0x78, 0x47, 0x4e },
{ 0x65, 0xa0, 0xee, 0x39, 0xd1, 0xf7, 0x38, 0x83, 0xf7, 0x5e, 0xe9, 0x37, 0xe4, 0x2c, 0x3a, 0xbd, 0x21, 0x97, 0xb2, 0x26, 0x1, 0x13, 0xf8, 0x6f, 0xa3, 0x44, 0xed, 0xd1, 0xef, 0x9f, 0xde, 0xe7 },
{ 0x8b, 0xa0, 0xdf, 0x15, 0x76, 0x25, 0x92, 0xd9, 0x3c, 0x85, 0xf7, 0xf6, 0x12, 0xdc, 0x42, 0xbe, 0xd8, 0xa7, 0xec, 0x7c, 0xab, 0x27, 0xb0, 0x7e, 0x53, 0x8d, 0x7d, 0xda, 0xaa, 0x3e, 0xa8, 0xde },
{ 0xaa, 0x25, 0xce, 0x93, 0xbd, 0x2, 0x69, 0xd8, 0x5a, 0xf6, 0x43, 0xfd, 0x1a, 0x73, 0x8, 0xf9, 0xc0, 0x5f, 0xef, 0xda, 0x17, 0x4a, 0x19, 0xa5, 0x97, 0x4d, 0x66, 0x33, 0x4c, 0xfd, 0x21, 0x6a },
{ 0x35, 0xb4, 0x98, 0x31, 0xdb, 0x41, 0x15, 0x70, 0xea, 0x1e, 0xf, 0xbb, 0xed, 0xcd, 0x54, 0x9b, 0x9a, 0xd0, 0x63, 0xa1, 0x51, 0x97, 0x40, 0x72, 0xf6, 0x75, 0x9d, 0xbf, 0x91, 0x47, 0x6f, 0xe2 } };
/* 1344 bytes */
__constant__ static __align__(16) uint32_t c_E8_bslice32[42][8] = {
// Round 0 (Function0)
{ 0xa2ded572, 0x90d6ab81, 0x67f815df, 0xf6875a4d, 0x0a15847b, 0xc54f9f4e, 0x571523b7, 0x402bd1c3 },
{ 0xe03a98ea, 0xb4960266, 0x9cfa455c, 0x8a53bbf2, 0x99d2c503, 0x1a1456b5, 0x9a99b266, 0x31a2db88 }, // 1
{ 0x5c5aa303, 0x8019051c, 0xdb0e199a, 0x1d959e84, 0x0ab23f40, 0xadeb336f, 0x1044c187, 0xdccde75e }, // 2
{ 0x9213ba10, 0x39812c0a, 0x416bbf02, 0x5078aa37, 0x156578dc, 0xd2bf1a3f, 0xd027bbf7, 0xd3910041 }, // 3
{ 0x0d5a2d42, 0x0ba75c18, 0x907eccf6, 0xac442bc7, 0x9c9f62dd, 0xd665dfd1, 0xce97c092, 0x23fcc663 }, // 4
{ 0x036c6e97, 0xbb03f1ee, 0x1ab8e09e, 0xfa618e5d, 0x7e450521, 0xb29796fd, 0xa8ec6c44, 0x97818394 }, // 5
{ 0x37858e4a, 0x8173fe8a, 0x2f3003db, 0x6c69b8f8, 0x2d8d672a, 0x4672c78a, 0x956a9ffb, 0x14427fc0 }, // 6
// Round 7 (Function0)
{ 0x8f15f4c5, 0xb775de52, 0xc45ec7bd, 0xbc88e4ae, 0xa76f4475, 0x1e00b882, 0x80bb118f, 0xf4a3a698 },
{ 0x338ff48e, 0x20edf1b6, 0x1563a3a9, 0xfde05a7c, 0x24565faa, 0x5ae9ca36, 0x89f9b7d5, 0x362c4206 },
{ 0x433529ce, 0x591ff5d0, 0x3d98fe4e, 0x86814e6f, 0x74f93a53, 0x81ad9d0e, 0xa74b9a73, 0x9f5ad8af },
{ 0x670605a7, 0x26077447, 0x6a6234ee, 0x3f1080c6, 0xbe280b8b, 0x6f7ea0e0, 0x2717b96e, 0x7b487ec6 },
{ 0xa50a550d, 0x81727686, 0xc0a4f84a, 0xd48d6050, 0x9fe7e391, 0x415a9e7e, 0x9ef18e97, 0x62b0e5f3 },
{ 0xec1f9ffc, 0xf594d74f, 0x7a205440, 0xd895fa9d, 0x001ae4e3, 0x117e2e55, 0x84c9f4ce, 0xa554c324 },
{ 0x2872df5b, 0xef7c8905, 0x286efebd, 0x2ed349ee, 0xe27ff578, 0x85937e44, 0xb2c4a50f, 0x7f5928eb },
// Round 14 (Function0)
{ 0x37695f70, 0x04771bc7, 0x4a3124b3, 0xe720b951, 0xf128865e, 0xe843fe74, 0x65e4d61d, 0x8a87d423 },
{ 0xa3e8297d, 0xfb301b1d, 0xf2947692, 0xe01bdc5b, 0x097acbdd, 0x4f4924da, 0xc1d9309b, 0xbf829cf2 },
{ 0x31bae7a4, 0x32fcae3b, 0xffbf70b4, 0x39d3bb53, 0x0544320d, 0xc1c39f45, 0x48bcf8de, 0xa08b29e0 },
{ 0xfd05c9e5, 0x01b771a2, 0x0f09aef7, 0x95ed44e3, 0x12347094, 0x368e3be9, 0x34f19042, 0x4a982f4f },
{ 0x631d4088, 0xf14abb7e, 0x15f66ca0, 0x30c60ae2, 0x4b44c147, 0xc5b67046, 0xffaf5287, 0xe68c6ecc },
{ 0x56a4d5a4, 0x45ce5773, 0x00ca4fbd, 0xadd16430, 0x4b849dda, 0x68cea6e8, 0xae183ec8, 0x67255c14 },
{ 0xf28cdaa3, 0x20b2601f, 0x16e10ecb, 0x7b846fc2, 0x5806e933, 0x7facced1, 0x9a99949a, 0x1885d1a0 },
// Round 21 (Function0)
{ 0xa15b5932, 0x67633d9f, 0xd319dd8d, 0xba6b04e4, 0xc01c9a50, 0xab19caf6, 0x46b4a5aa, 0x7eee560b },
{ 0xea79b11f, 0x5aac571d, 0x742128a9, 0x76d35075, 0x35f7bde9, 0xfec2463a, 0xee51363b, 0x01707da3 },
{ 0xafc135f7, 0x15638341, 0x42d8a498, 0xa8db3aea, 0x20eced78, 0x4d3bc3fa, 0x79676b9e, 0x832c8332 },
{ 0x1f3b40a7, 0x6c4e3ee7, 0xf347271c, 0xfd4f21d2, 0x34f04059, 0x398dfdb8, 0x9a762db7, 0xef5957dc },
{ 0x490c9b8d, 0xd0ae3b7d, 0xdaeb492b, 0x84558d7a, 0x49d7a25b, 0xf0e9a5f5, 0x0d70f368, 0x658ef8e4 },
{ 0xf4a2b8a0, 0x92946891, 0x533b1036, 0x4f88e856, 0x9e07a80c, 0x555cb05b, 0x5aec3e75, 0x4cbcbaf8 },
{ 0x993bbbe3, 0x28acae64, 0x7b9487f3, 0x6db334dc, 0xd6f4da75, 0x50a5346c, 0x5d1c6b72, 0x71db28b8 },
// Round 28 (Function0)
{ 0xf2e261f8, 0xf1bcac1c, 0x2a518d10, 0xa23fce43, 0x3364dbe3, 0x3cd1bb67, 0xfc75dd59, 0xb043e802 },
{ 0xca5b0a33, 0xc3943b92, 0x75a12988, 0x1e4d790e, 0x4d19347f, 0xd7757479, 0x5c5316b4, 0x3fafeeb6 },
{ 0xf7d4a8ea, 0x5324a326, 0x21391abe, 0xd23c32ba, 0x097ef45c, 0x4a17a344, 0x5127234c, 0xadd5a66d },
{ 0xa63e1db5, 0xa17cf84c, 0x08c9f2af, 0x4d608672, 0x983d5983, 0xcc3ee246, 0x563c6b91, 0xf6c76e08 },
{ 0xb333982f, 0xe8b6f406, 0x5e76bcb1, 0x36d4c1be, 0xa566d62b, 0x1582ee74, 0x2ae6c4ef, 0x6321efbc },
{ 0x0d4ec1fd, 0x1614c17e, 0x69c953f4, 0x16fae006, 0xc45a7da7, 0x3daf907e, 0x26585806, 0x3f9d6328 },
{ 0xe3f2c9d2, 0x16512a74, 0x0cd29b00, 0x9832e0f2, 0x30ceaa5f, 0xd830eb0d, 0x300cd4b7, 0x9af8cee3 },
// Round 35 (Function0)
{ 0x7b9ec54b, 0x574d239b, 0x9279f1b5, 0x316796e6, 0x6ee651ff, 0xf3a6e6cc, 0xd3688604, 0x05750a17 },
{ 0xd98176b1, 0xb3cb2bf4, 0xce6c3213, 0x47154778, 0x8452173c, 0x825446ff, 0x62a205f8, 0x486a9323 },
{ 0x0758df38, 0x442e7031, 0x65655e4e, 0x86ca0bd0, 0x897cfcf2, 0xa20940f0, 0x8e5086fc, 0x4e477830 },
{ 0x39eea065, 0x26b29721, 0x8338f7d1, 0x6ff81301, 0x37e95ef7, 0xd1ed44a3, 0xbd3a2ce4, 0xe7de9fef },
{ 0x15dfa08b, 0x7ceca7d8, 0xd9922576, 0x7eb027ab, 0xf6f7853c, 0xda7d8d53, 0xbe42dc12, 0xdea83eaa },
{ 0x93ce25aa, 0xdaef5fc0, 0xd86902bd, 0xa5194a17, 0xfd43f65a, 0x33664d97, 0xf908731a, 0x6a21fd4c },
{ 0x3198b435, 0xa163d09a, 0x701541db, 0x72409751, 0xbb0f1eea, 0xbf9d75f6, 0x9b54cded, 0xe26f4791 }
// 42 rounds...
};
#define SWAP4(x,y)\
y = (x & 0xf0f0f0f0UL); \
x = (x ^ y); \
y = (y >> 4); \
x = (x << 4); \
x= x | y;
#define SWAP2(x,y)\
y = (x & 0xccccccccUL); \
x = (x ^ y); \
y = (y >> 2); \
x = (x << 2); \
x= x | y;
#define SWAP1(x,y)\
y = (x & 0xaaaaaaaaUL); \
x = (x ^ y); \
y = (y >> 1); \
x = x + x; \
x= x | y;
/*swapping bits 16i||16i+1||......||16i+7 with bits 16i+8||16i+9||......||16i+15 of 32-bit x*/
//#define SWAP8(x) (x) = ((((x) & 0x00ff00ffUL) << 8) | (((x) & 0xff00ff00UL) >> 8));
#define SWAP8(x) (x) = __byte_perm(x, x, 0x2301);
/*swapping bits 32i||32i+1||......||32i+15 with bits 32i+16||32i+17||......||32i+31 of 32-bit x*/
//#define SWAP16(x) (x) = ((((x) & 0x0000ffffUL) << 16) | (((x) & 0xffff0000UL) >> 16));
#define SWAP16(x) (x) = __byte_perm(x, x, 0x1032);
/*swapping bits 16i||16i+1||......||16i+7 with bits 16i+8||16i+9||......||16i+15 of 32-bit x*/
//#define SWAP8(x) (x) = ((((x) & 0x00ff00ffUL) << 8) | (((x) & 0xff00ff00UL) >> 8));
#define SWAP8(x) (x) = __byte_perm(x, x, 0x2301);
/*
__device__ __forceinline__
static void SWAP4(uint32_t &x) {
uint32_t y = x & 0xF0F0F0F0;
x = (x ^ y) << 4;
x |= y >> 4;
}
__device__ __forceinline__
static void SWAP2(uint32_t &x) {
uint32_t y = (x & 0xCCCCCCCC);
x = (x ^ y) << 2;
x |= y >> 2;
}
__device__ __forceinline__
static void SWAP1(uint32_t &x) {
uint32_t y = (x & 0xAAAAAAAA);
x = (x ^ y) << 1;
x |= y >> 1;
}
*/
__device__ __forceinline__
static void SWAP4x4(uint32_t *x) {
#pragma nounroll
// y is used as tmp register too
for (uint32_t y=0; y<4; y++, ++x) {
asm("and.b32 %1, %0, 0xF0F0F0F0;"
"xor.b32 %0, %0, %1;"
"shr.b32 %1, %1, 4;"
"vshl.u32.u32.u32.clamp.add %0, %0, 4, %1;\n\t"
: "+r"(*x) : "r"(y));
}
}
__device__ __forceinline__
static void SWAP2x4(uint32_t *x) {
#pragma nounroll
// y is used as tmp register too
for (uint32_t y=0; y<4; y++, ++x) {
asm("and.b32 %1, %0, 0xCCCCCCCC;"
"xor.b32 %0, %0, %1;"
"shr.b32 %1, %1, 2;"
"vshl.u32.u32.u32.clamp.add %0, %0, 2, %1;\n\t"
: "+r"(*x) : "r"(y));
}
}
__device__ __forceinline__
static void SWAP1x4(uint32_t *x) {
#pragma nounroll
// y is used as tmp register too
for (uint32_t y=0; y<4; y++, ++x) {
asm("and.b32 %1, %0, 0xAAAAAAAA;"
"xor.b32 %0, %0, %1;"
"shr.b32 %1, %1, 1;"
"vshl.u32.u32.u32.clamp.add %0, %0, 1, %1;\n\t"
: "+r"(*x) : "r"(y));
}
}
/* The MDS transform */
/*The MDS transform*/
#define L(m0,m1,m2,m3,m4,m5,m6,m7) \
m4 ^= m1; \
m5 ^= m2; \
m6 ^= m0 ^ m3; \
m7 ^= m0; \
m0 ^= m5; \
m1 ^= m6; \
m2 ^= m4 ^ m7; \
m3 ^= m4;
(m4) ^= (m1); \
(m5) ^= (m2); \
(m6) ^= (m0) ^ (m3); \
(m7) ^= (m0); \
(m0) ^= (m5); \
(m1) ^= (m6); \
(m2) ^= (m4) ^ (m7); \
(m3) ^= (m4);
/* The Sbox */
#define Sbox(m0, m1, m2, m3, cc) \
/*The Sbox*/
#define Sbox(m0,m1,m2,m3,cc) \
m3 = ~(m3); \
m0 ^= (~(m2)) & cc; \
temp0 = cc ^ (m0 & m1); \
m0 ^= m2 & m3; \
m3 ^= (~(m1)) & m2; \
m1 ^= m0 & m2; \
m2 ^= m0 & (~(m3)); \
m0 ^= m1 | m3; \
m3 ^= m1 & m2; \
m1 ^= temp0 & m0; \
m0 ^= ((~(m2)) & (cc)); \
temp0 = (cc) ^ ((m0) & (m1));\
m0 ^= ((m2) & (m3)); \
m3 ^= ((~(m1)) & (m2)); \
m1 ^= ((m0) & (m2)); \
m2 ^= ((m0) & (~(m3))); \
m0 ^= ((m1) | (m3)); \
m3 ^= ((m1) & (m2)); \
m1 ^= (temp0 & (m0)); \
m2 ^= temp0;
__device__ __forceinline__
static void Sbox_and_MDS_layer(uint32_t x[8][4], const int rnd)
static __device__ __forceinline__ void Sbox_and_MDS_layer(uint32_t x[8][4], uint32_t roundnumber)
{
uint2* cc = (uint2*) &c_E8_bslice32[rnd];
// Sbox and MDS layer
#pragma unroll
for (int i = 0; i < 4; i++, ++cc) {
uint32_t temp0;
Sbox(x[0][i], x[2][i], x[4][i], x[6][i], cc->x);
Sbox(x[1][i], x[3][i], x[5][i], x[7][i], cc->y);
uint32_t temp0;
uint32_t cc0, cc1;
//Sbox and MDS layer
#pragma unroll 4
for (int i = 0; i < 4; i++) {
cc0 = ((uint32_t*)c_E8_bitslice_roundconstant[roundnumber])[i];
cc1 = ((uint32_t*)c_E8_bitslice_roundconstant[roundnumber])[i + 4];
Sbox(x[0][i], x[2][i], x[4][i], x[6][i], cc0);
Sbox(x[1][i], x[3][i], x[5][i], x[7][i], cc1);
L(x[0][i], x[2][i], x[4][i], x[6][i], x[1][i], x[3][i], x[5][i], x[7][i]);
}
}
__device__ __forceinline__
static void RoundFunction0(uint32_t x[8][4], const int rnd)
static __device__ __forceinline__ void RoundFunction0(uint32_t x[8][4], uint32_t roundnumber)
{
Sbox_and_MDS_layer(x, rnd + 0); // 0, 7, 14 .. 35
#pragma unroll 4
for (int j = 1; j < 8; j += 2) { // 1, 3, 5, 7 (Even)
SWAP1x4(x[j]);
// SWAP1(x[j][0]); SWAP1(x[j][1]); SWAP1(x[j][2]); SWAP1(x[j][3]);
}
}
Sbox_and_MDS_layer(x, roundnumber);
__device__ __forceinline__
static void RoundFunction1(uint32_t x[8][4], const int rnd)
{
Sbox_and_MDS_layer(x, rnd + 1);
#pragma unroll 4
for (int j = 1; j < 8; j += 2) {
SWAP2x4(x[j]);
// SWAP2(x[j][0]); SWAP2(x[j][1]); SWAP2(x[j][2]); SWAP2(x[j][3]);
}
}
__device__ __forceinline__
static void RoundFunction2(uint32_t x[8][4], const int rnd)
{
Sbox_and_MDS_layer(x, rnd + 2);
#pragma unroll 4
for (int j = 1; j < 8; j += 2) {
SWAP4x4(x[j]);
// SWAP4(x[j][0]); SWAP4(x[j][1]); SWAP4(x[j][2]); SWAP4(x[j][3]);
}
}
__device__ __forceinline__
static void RoundFunction3(uint32_t x[8][4], const int rnd)
{
Sbox_and_MDS_layer(x, rnd + 3);
//uint32_t* xj = x[j];
#pragma unroll 4
for (int j = 1; j < 8; j += 2) {
SWAP8(x[j][0]);
SWAP8(x[j][1]);
SWAP8(x[j][2]);
SWAP8(x[j][3]);
}
}
__device__ __forceinline__
static void RoundFunction4(uint32_t x[8][4], const int rnd)
{
Sbox_and_MDS_layer(x, rnd + 4);
#pragma unroll 4
for (int j = 1; j < 8; j += 2)
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
//uint32_t* xj = x[j];
#pragma unroll
for (int i = 0; i < 4; i++)
SWAP16(x[j][i]);
uint32_t y;
SWAP1(x[j][0], y);
SWAP1(x[j][1], y);
SWAP1(x[j][2], y);
SWAP1(x[j][3], y);
}
}
__device__ __forceinline__
static void RoundFunction5(uint32_t x[8][4], const int rnd)
static __device__ __forceinline__ void RoundFunction1(uint32_t x[8][4], uint32_t roundnumber)
{
Sbox_and_MDS_layer(x, rnd + 5);
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j += 2)
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
xchg(x[j][0], x[j][1]);
xchg(x[j][2], x[j][3]);
uint32_t y;
SWAP2(x[j][0], y);
SWAP2(x[j][1], y);
SWAP2(x[j][2], y);
SWAP2(x[j][3], y);
}
}
__device__ __forceinline__
static void RoundFunction6(uint32_t x[8][4], const int rnd)
static __device__ __forceinline__ void RoundFunction2(uint32_t x[8][4], uint32_t roundnumber)
{
Sbox_and_MDS_layer(x, rnd + 6);
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j += 2)
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
xchg(x[j][0], x[j][2]);
xchg(x[j][1], x[j][3]);
uint32_t y;
SWAP4(x[j][0], y);
SWAP4(x[j][1], y);
SWAP4(x[j][2], y);
SWAP4(x[j][3], y);
}
}
/* The bijective function E8, in bitslice form */
__device__
static void E8(uint32_t x[8][4])
static __device__ __forceinline__ void RoundFunction3(uint32_t x[8][4], uint32_t roundnumber)
{
/* perform 6 loops of 7 rounds */
for (int r = 0; r < 42; r += 7)
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
RoundFunction0(x, r);
RoundFunction1(x, r);
RoundFunction2(x, r);
RoundFunction3(x, r);
RoundFunction4(x, r);
RoundFunction5(x, r);
RoundFunction6(x, r);
#pragma unroll 4
for (int i = 0; i < 4; i++) SWAP8(x[j][i]);
}
}
static __device__ __forceinline__ void RoundFunction4(uint32_t x[8][4], uint32_t roundnumber)
{
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
#pragma unroll 4
for (int i = 0; i < 4; i++) SWAP16(x[j][i]);
}
}
static __device__ __forceinline__ void RoundFunction5(uint32_t x[8][4], uint32_t roundnumber)
{
uint32_t temp0;
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
#pragma unroll 2
for (int i = 0; i < 4; i = i + 2) {
temp0 = x[j][i]; x[j][i] = x[j][i + 1]; x[j][i + 1] = temp0;
}
}
}
static __device__ __forceinline__ void RoundFunction6(uint32_t x[8][4], uint32_t roundnumber)
{
uint32_t temp0;
Sbox_and_MDS_layer(x, roundnumber);
#pragma unroll 4
for (int j = 1; j < 8; j = j + 2)
{
#pragma unroll 2
for (int i = 0; i < 2; i++) {
temp0 = x[j][i]; x[j][i] = x[j][i + 2]; x[j][i + 2] = temp0;
}
}
}
/*The bijective function E8, in bitslice form */
static __device__ __forceinline__ void E8(uint32_t x[8][4])
{
/*perform 6 rounds*/
//#pragma unroll 6
for (int i = 0; i < 42; i += 7)
{
RoundFunction0(x, i);
RoundFunction1(x, i + 1);
RoundFunction2(x, i + 2);
RoundFunction3(x, i + 3);
RoundFunction4(x, i + 4);
RoundFunction5(x, i + 5);
RoundFunction6(x, i + 6);
}
}
// Die Hash-Funktion
__global__ __launch_bounds__(256, 4)
void quark_jh512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t* g_hash, const uint32_t *const __restrict__ g_nonceVector)
void quark_jh512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t *const __restrict__ g_hash, const uint32_t *const __restrict__ g_nonceVector)
{
const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
const uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
uint32_t hashPosition = nounce - startNounce;
uint32_t *Hash = &g_hash[hashPosition * 16U];
uint32_t x[8][4] = { /* init */
const uint32_t hashPosition = nounce - startNounce;
uint32_t *const Hash = &g_hash[hashPosition * 16U];
uint32_t x[8][4] = {
{ 0x964bd16f, 0x17aa003e, 0x052e6a63, 0x43d5157a },
{ 0x8d5e228a, 0x0bef970c, 0x591234e9, 0x61c3b3f2 },
{ 0xc1a01d89, 0x1e806f53, 0x6b05a92a, 0x806d2bea },
@@ -292,22 +250,21 @@ void quark_jh512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t* g
{ 0xdffcc2e3, 0xfb1785e6, 0x78465a54, 0x4bdd8ccc }
};
#pragma unroll
#pragma unroll 16
for (int i = 0; i < 16; i++)
x[i/4][i & 3] ^= Hash[i];
x[i >> 2][i & 3] ^= Hash[i];
E8(x);
#pragma unroll
for (uint8_t i = 0; i < 16; i++)
x[(i+16)/4][(i+16) & 3] ^= Hash[i];
#pragma unroll 16
for (int i = 0; i < 16; i++)
x[(16 + i) >> 2][(16 + i) & 3] ^= Hash[i];
x[0][0] ^= 0x80U;
x[3][3] ^= 0x00020000U;
x[0 >> 2][0 & 3] ^= 0x80;
x[15 >> 2][15 & 3] ^= 0x00020000;
E8(x);
x[4][0] ^= 0x80U;
x[7][3] ^= 0x00020000U;
x[(16 + 0) >> 2][(16 + 0) & 3] ^= 0x80;
x[(16 + 15) >> 2][(16 + 15) & 3] ^= 0x00020000;
Hash[0] = x[4][0];
Hash[1] = x[4][1];
+7 -4
View File
@@ -1,6 +1,5 @@
#include <stdio.h>
#include <stdio.h>
#include <memory.h>
#include <sys/types.h> // off_t
#include "cuda_helper.h"
@@ -120,8 +119,12 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
#if USE_SHUFFLE
const uint32_t warpBlockID = (thread + 15)>>4; // aufrunden auf volle Warp-Blöcke
const int warpID = threadIdx.x & 0x0F; // 16 warps
const int warpBlockID = (thread + 15)>>4; // aufrunden auf volle Warp-Blöcke
const int maxHashPosition = thread<<3;
#endif
#if USE_SHUFFLE
if (warpBlockID < ( (threads+15)>>4 ))
#else
if (thread < threads)
@@ -129,7 +132,7 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t
{
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
int hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[hashPosition<<3]; // hashPosition * 8
// 128 Bytes
+10 -9
View File
@@ -2,7 +2,6 @@
#include <stdio.h>
#include <memory.h>
#include <sys/types.h> // off_t
#include "cuda_helper.h"
@@ -29,19 +28,19 @@ void quark_groestl512_gpu_hash_64_quad(uint32_t threads, uint32_t startNounce, u
uint32_t state[8];
uint32_t nounce = g_nonceVector ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
uint32_t *pHash = &g_hash[hashPosition << 4];
int hashPosition = nounce - startNounce;
uint32_t *inpHash = &g_hash[hashPosition << 4];
const uint32_t thr = threadIdx.x % THF;
const uint16_t thr = threadIdx.x % THF;
#pragma unroll
for(int k=0;k<4;k++) message[k] = pHash[thr + (k * THF)];
for(int k=0;k<4;k++) message[k] = inpHash[(k * THF) + thr];
#pragma unroll
for(int k=4;k<8;k++) message[k] = 0;
if (thr == 0) message[4] = 0x80U;
if (thr == 3) message[7] = 0x01000000U;
if (thr == 0) message[4] = 0x80;
if (thr == 3) message[7] = 0x01000000;
uint32_t msgBitsliced[8];
to_bitslice_quad(message, msgBitsliced);
@@ -49,13 +48,15 @@ void quark_groestl512_gpu_hash_64_quad(uint32_t threads, uint32_t startNounce, u
groestl512_progressMessage_quad(state, msgBitsliced);
// Nur der erste von jeweils 4 Threads bekommt das Ergebns-Hash
uint32_t *outpHash = inpHash;
uint32_t hash[16];
from_bitslice_quad(state, hash);
// uint4 = 4x4 uint32_t = 16 bytes
if (thr == 0) {
uint4 *phash = (uint4*) hash;
uint4 *outpt = (uint4*) pHash;
uint4 *outpt = (uint4*) outpHash; /* var kept for hash align */
outpt[0] = phash[0];
outpt[1] = phash[1];
outpt[2] = phash[2];
@@ -84,7 +85,7 @@ __global__ void __launch_bounds__(TPB, THF)
uint32_t nounce = g_nonceVector ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
int hashPosition = nounce - startNounce;
uint32_t * inpHash = &g_hash[hashPosition<<4];
const uint16_t thr = threadIdx.x % THF;
+3 -3
View File
@@ -230,8 +230,8 @@ void quark_groestl512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
uint32_t *inpHash = &g_hash[hashPosition * 16];
int hashPosition = nounce - startNounce;
uint32_t *inpHash = &g_hash[16 * hashPosition];
#pragma unroll 16
for(int k=0; k<16; k++)
@@ -266,7 +266,7 @@ void quark_groestl512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32
for(int u=0;u<32;u++) state[u] ^= message[u];
// Erzeugten Hash rausschreiben
uint32_t *outpHash = &g_hash[hashPosition * 16];
uint32_t *outpHash = &g_hash[16 * hashPosition];
#pragma unroll 16
for(int k=0;k<16;k++) outpHash[k] = state[k+16];
+5 -6
View File
@@ -1,6 +1,5 @@
#include <stdio.h>
#include <memory.h>
#include <sys/types.h> // off_t
#include "cuda_helper.h"
@@ -102,8 +101,8 @@ void quark_keccak512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_
{
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[hashPosition * 8];
int hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[8 * hashPosition];
uint2 keccak_gpu_state[25];
for (int i = 0; i<8; i++) {
@@ -201,8 +200,8 @@ void quark_keccak512_gpu_hash_64_v30(uint32_t threads, uint32_t startNounce, uin
{
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
off_t hashPosition = nounce - startNounce;
uint32_t *inpHash = (uint32_t*)&g_hash[hashPosition * 8];
int hashPosition = nounce - startNounce;
uint32_t *inpHash = (uint32_t*)&g_hash[8 * hashPosition];
uint32_t message[18];
#pragma unroll 16
@@ -225,7 +224,7 @@ void quark_keccak512_gpu_hash_64_v30(uint32_t threads, uint32_t startNounce, uin
U64TO32_LE((&hash[i/4]), keccak_gpu_state[i / 8]);
}
uint32_t *outpHash = (uint32_t*)&g_hash[hashPosition * 8];
uint32_t *outpHash = (uint32_t*)&g_hash[8 * hashPosition];
#pragma unroll 16
for(int i=0; i<16; i++)
outpHash[i] = hash[i];
+6 -16
View File
@@ -418,19 +418,13 @@ void quark_skein512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t
h7 = vectorize(0xAE18A40B660FCC33ull);
uint2 p[8];
// 1st Round -> etype = 480, ptr = 64, bcount = 0, data = msg
// 1. Runde -> etype = 480, ptr = 64, bcount = 0, data = msg
#pragma unroll 8
for (int i = 0; i < 8; i++)
p[i] = vectorize(inpHash[i]);
t0 = vectorize(64); // ptr
// t1 = vectorize(480ull << 55); // etype
t1 = vectorize(0xf000000000000000ULL);
//#if CUDA_VERSION >= 7000
// doesnt really affect x11 perfs.
__threadfence();
//#endif
t1 = vectorize(480ull << 55); // etype
TFBIG_KINIT_UI2(h0, h1, h2, h3, h4, h5, h6, h7, h8, t0, t1, t2);
TFBIG_4e_UI2(0);
TFBIG_4o_UI2(1);
@@ -464,12 +458,10 @@ void quark_skein512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint64_t
// 2. Runde -> etype = 510, ptr = 8, bcount = 0, data = 0
#pragma unroll 8
for(int i=0; i<8; i++)
p[i] = vectorize(0);
p[i] = vectorize(0ull);
t0 = vectorize(8); // ptr
//t1 = vectorize(510ull << 55); // etype
t1 = vectorize(0xff00000000000000ULL);
t1 = vectorize(510ull << 55); // etype
TFBIG_KINIT_UI2(h0, h1, h2, h3, h4, h5, h6, h7, h8, t0, t1, t2);
TFBIG_4e_UI2(0);
TFBIG_4o_UI2(1);
@@ -512,7 +504,7 @@ void quark_skein512_gpu_hash_64_sm3(uint32_t threads, uint32_t startNounce, uint
uint32_t nounce = (g_nonceVector != NULL) ? g_nonceVector[thread] : (startNounce + thread);
uint32_t hashPosition = nounce - startNounce;
int hashPosition = nounce - startNounce;
uint64_t *inpHash = &g_hash[hashPosition * 8];
// Init
@@ -531,9 +523,7 @@ void quark_skein512_gpu_hash_64_sm3(uint32_t threads, uint32_t startNounce, uint
p[i] = inpHash[i];
t0 = 64; // ptr
// t1 = 480ull << 55; // etype
t1 = 0xf000000000000000ULL;
t1 = 480ull << 55; // etype
TFBIG_KINIT(h0, h1, h2, h3, h4, h5, h6, h7, h8, t0, t1, t2);
TFBIG_4e(0);
TFBIG_4o(1);
+16 -13
View File
@@ -19,6 +19,8 @@ static uint32_t *d_hash[MAX_GPUS];
extern void qubit_luffa512_cpu_init(int thr_id, uint32_t threads);
extern void qubit_luffa512_cpu_setBlock_80(void *pdata);
extern void qubit_luffa512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
extern void qubit_luffa512_cpufinal_setBlock_80(void *pdata, const void *ptarget);
extern uint32_t qubit_luffa512_cpu_finalhash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
extern void x11_cubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_cubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -28,13 +30,13 @@ extern void x11_echo512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t start
extern "C" void deephash(void *state, const void *input)
{
uint8_t _ALIGN(64) hash[64];
// luffa-80 cubehash-64 echo-64
// luffa1-cubehash2-shavite3-simd4-echo5
sph_luffa512_context ctx_luffa;
sph_cubehash512_context ctx_cubehash;
sph_echo512_context ctx_echo;
uint8_t hash[64];
sph_luffa512_init(&ctx_luffa);
sph_luffa512 (&ctx_luffa, input, 80);
sph_luffa512_close(&ctx_luffa, (void*) hash);
@@ -52,13 +54,14 @@ extern "C" void deephash(void *state, const void *input)
static bool init[MAX_GPUS] = { 0 };
extern "C" int scanhash_deep(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
extern "C" int scanhash_deep(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done)
{
uint32_t _ALIGN(64) endiandata[20];
const uint32_t first_nonce = pdata[19];
uint32_t endiandata[20];
uint32_t throughput = device_intensity(thr_id, __func__, 1U << 19); // 256*256*8
throughput = min(throughput, (max_nonce - first_nonce));
throughput = min(throughput, (max_nonce - first_nonce));
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x0000f;
@@ -66,8 +69,7 @@ extern "C" int scanhash_deep(int thr_id, uint32_t *pdata, const uint32_t *ptarge
if (!init[thr_id])
{
cudaSetDevice(device_map[thr_id]);
CUDA_SAFE_CALL(cudaMalloc(&d_hash[thr_id], throughput * 64));
cudaMalloc(&d_hash[thr_id], 16 * sizeof(uint32_t) * throughput);
qubit_luffa512_cpu_init(thr_id, throughput);
x11_cubehash512_cpu_init(thr_id, throughput);
@@ -78,10 +80,10 @@ extern "C" int scanhash_deep(int thr_id, uint32_t *pdata, const uint32_t *ptarge
init[thr_id] = true;
}
for (int k=0; k < 19; k++)
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
qubit_luffa512_cpu_setBlock_80((void*)endiandata);
qubit_luffa512_cpufinal_setBlock_80((void*)endiandata,ptarget);
cuda_check_cpu_setTarget(ptarget);
do {
@@ -94,11 +96,12 @@ extern "C" int scanhash_deep(int thr_id, uint32_t *pdata, const uint32_t *ptarge
uint32_t foundNonce = cuda_check_hash(thr_id, throughput, pdata[19], d_hash[thr_id]);
if (foundNonce != UINT32_MAX)
{
uint32_t _ALIGN(64) vhash64[8];
const uint32_t Htarg = ptarget[7];
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
deephash(vhash64, endiandata);
if (vhash64[7] <= ptarget[7] && fulltest(vhash64, ptarget)) {
if (vhash64[7] <= Htarg && fulltest(vhash64, ptarget)) {
int res = 1;
uint32_t secNonce = cuda_check_hash_suppl(thr_id, throughput, pdata[19], d_hash[thr_id], 1);
*hashes_done = pdata[19] - first_nonce + throughput;
+23 -22
View File
@@ -1,5 +1,6 @@
/*
* luffa 80 algo (Introduced by Doomcoin)
* qubit algorithm
*
*/
extern "C" {
#include "sph/sph_luffa.h"
@@ -14,13 +15,16 @@ static uint32_t *d_hash[MAX_GPUS];
extern void qubit_luffa512_cpu_init(int thr_id, uint32_t threads);
extern void qubit_luffa512_cpu_setBlock_80(void *pdata);
extern void qubit_luffa512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
extern void qubit_luffa512_cpufinal_setBlock_80(void *pdata, const void *ptarget);
extern uint32_t qubit_luffa512_cpu_finalhash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
extern "C" void luffa_hash(void *state, const void *input)
extern void doomhash(void *state, const void *input)
{
uint8_t _ALIGN(64) hash[64];
// luffa512
sph_luffa512_context ctx_luffa;
uint8_t hash[64];
sph_luffa512_init(&ctx_luffa);
sph_luffa512 (&ctx_luffa, input, 80);
sph_luffa512_close(&ctx_luffa, (void*) hash);
@@ -30,11 +34,12 @@ extern "C" void luffa_hash(void *state, const void *input)
static bool init[MAX_GPUS] = { 0 };
extern "C" int scanhash_luffa(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
extern "C" int scanhash_doom(int thr_id, uint32_t *pdata,
const uint32_t *ptarget, uint32_t max_nonce,
unsigned long *hashes_done)
{
uint32_t _ALIGN(64) endiandata[20];
const uint32_t first_nonce = pdata[19];
uint32_t endiandata[20];
uint32_t throughput = device_intensity(thr_id, __func__, 1U << 22); // 256*256*8*8
throughput = min(throughput, max_nonce - first_nonce);
@@ -45,35 +50,31 @@ extern "C" int scanhash_luffa(int thr_id, uint32_t *pdata, const uint32_t *ptarg
{
cudaSetDevice(device_map[thr_id]);
CUDA_SAFE_CALL(cudaMalloc(&d_hash[thr_id], throughput * 64));
CUDA_SAFE_CALL(cudaMalloc(&d_hash[thr_id], 16 * sizeof(uint32_t) * throughput));
qubit_luffa512_cpu_init(thr_id, throughput);
cuda_check_cpu_init(thr_id, throughput);
qubit_luffa512_cpu_init(thr_id, (int) throughput);
init[thr_id] = true;
}
for (int k=0; k < 19; k++)
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
qubit_luffa512_cpu_setBlock_80((void*)endiandata);
cuda_check_cpu_setTarget(ptarget);
qubit_luffa512_cpufinal_setBlock_80((void*)endiandata,ptarget);
do {
int order = 0;
*hashes_done = pdata[19] - first_nonce + throughput;
qubit_luffa512_cpu_hash_80(thr_id, (int) throughput, pdata[19], d_hash[thr_id], order++);
uint32_t foundNonce = cuda_check_hash(thr_id, throughput, pdata[19], d_hash[thr_id]);
uint32_t foundNonce = qubit_luffa512_cpu_finalhash_80(thr_id, (int) throughput, pdata[19], d_hash[thr_id], order++);
if (foundNonce != UINT32_MAX)
{
uint32_t _ALIGN(64) vhash64[8];
const uint32_t Htarg = ptarget[7];
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
luffa_hash(vhash64, endiandata);
doomhash(vhash64, endiandata);
if (vhash64[7] <= ptarget[7] && fulltest(vhash64, ptarget)) {
//*hashes_done = min(max_nonce - first_nonce, (uint64_t) pdata[19] - first_nonce + throughput);
if (vhash64[7] <= Htarg && fulltest(vhash64, ptarget)) {
*hashes_done = min(max_nonce - first_nonce, (uint64_t) pdata[19] - first_nonce + throughput);
pdata[19] = foundNonce;
return 1;
} else {
@@ -81,7 +82,7 @@ extern "C" int scanhash_luffa(int thr_id, uint32_t *pdata, const uint32_t *ptarg
}
}
if ((uint64_t) throughput + pdata[19] > max_nonce) {
if ((uint64_t) pdata[19] + throughput > max_nonce) {
// pdata[19] = max_nonce;
break;
}
+291 -290
View File
@@ -1,18 +1,49 @@
/*******************************************************************************
* luffa512 for 80-bytes input (with midstate precalc by klausT)
/*
* luffa_for_32.c
* Version 2.0 (Sep 15th 2009)
*
* Copyright (C) 2008-2009 Hitachi, Ltd. All rights reserved.
*
* Hitachi, Ltd. is the owner of this software and hereby grant
* the U.S. Government and any interested party the right to use
* this software for the purposes of the SHA-3 evaluation process,
* notwithstanding that this software is copyrighted.
*
* THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
* WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
* MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
* ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
* WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
* ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
* OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
*/
#include <stdio.h>
#include <stdint.h>
#include <memory.h>
#include "cuda_helper.h"
static __constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
static __constant__ uint32_t statebufferpre[8];
static __constant__ uint32_t statechainvpre[40];
typedef unsigned char BitSequence;
#define MULT2(a,j) {\
__constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
__constant__ uint32_t c_Target[8];
static uint32_t *h_resNounce[MAX_GPUS];
static uint32_t *d_resNounce[MAX_GPUS];
#define NBN 1 /* max results, could be 2, see blake32.cu */
#if NBN > 1
static uint32_t extra_results[2] = { UINT32_MAX, UINT32_MAX };
#endif
typedef struct {
uint32_t buffer[8]; /* Buffer to be hashed */
uint32_t chainv[40]; /* Chaining values */
} hashState;
#define BYTES_SWAP32(x) cuda_swab32(x)
#define MULT2(a,j)\
tmp = a[7+(8*j)];\
a[7+(8*j)] = a[6+(8*j)];\
a[6+(8*j)] = a[5+(8*j)];\
@@ -21,25 +52,22 @@ static __constant__ uint32_t statechainvpre[40];
a[3+(8*j)] = a[2+(8*j)] ^ tmp;\
a[2+(8*j)] = a[1+(8*j)];\
a[1+(8*j)] = a[0+(8*j)] ^ tmp;\
a[0+(8*j)] = tmp;\
}
a[0+(8*j)] = tmp;
#define TWEAK(a0,a1,a2,a3,j) { \
#define TWEAK(a0,a1,a2,a3,j)\
a0 = (a0<<(j))|(a0>>(32-j));\
a1 = (a1<<(j))|(a1>>(32-j));\
a2 = (a2<<(j))|(a2>>(32-j));\
a3 = (a3<<(j))|(a3>>(32-j));\
}
a3 = (a3<<(j))|(a3>>(32-j));
#define STEP(c0,c1) { \
#define STEP(c0,c1)\
SUBCRUMB(chainv[0],chainv[1],chainv[2],chainv[3],tmp);\
SUBCRUMB(chainv[5],chainv[6],chainv[7],chainv[4],tmp);\
MIXWORD(chainv[0],chainv[4]);\
MIXWORD(chainv[1],chainv[5]);\
MIXWORD(chainv[2],chainv[6]);\
MIXWORD(chainv[3],chainv[7]);\
ADD_CONSTANT(chainv[0],chainv[4],c0,c1);\
}
ADD_CONSTANT(chainv[0],chainv[4],c0,c1);
#define SUBCRUMB(a0,a1,a2,a3,a4)\
a4 = a0;\
@@ -76,7 +104,7 @@ static __constant__ uint32_t statechainvpre[40];
/* initial values of chaining variables */
__constant__ uint32_t c_IV[40];
static const uint32_t h_IV[40] = {
const uint32_t h2_IV[40] = {
0x6d251e69,0x44b051e0,0x4eaa6fb4,0xdbf78465,
0x6e292011,0x90152df4,0xee058139,0xdef610bb,
0xc3b44b95,0xd9d2f256,0x70eee9a0,0xde099fa3,
@@ -89,7 +117,7 @@ static const uint32_t h_IV[40] = {
0xf5df3999,0x0fc688f1,0xb07224cc,0x03e86cea};
__constant__ uint32_t c_CNS[80];
static const uint32_t h_CNS[80] = {
uint32_t h2_CNS[80] = {
0x303994a6,0xe0337818,0xc0e65299,0x441ba90d,
0x6cc33a12,0x7f34d442,0xdc56983e,0x9389217f,
0x1e00108f,0xe5a8bce6,0x7800423d,0x5274baf4,
@@ -114,305 +142,213 @@ static const uint32_t h_CNS[80] = {
/***************************************************/
__device__ __forceinline__
void rnd512(uint32_t *statebuffer, uint32_t *statechainv)
void rnd512(hashState *state)
{
int i,j;
uint32_t t[40];
uint32_t chainv[8];
uint32_t tmp;
#pragma unroll 8
for(i=0; i<8; i++) {
#pragma unroll 8
for(i=0;i<8;i++) {
t[i]=0;
#pragma unroll 5
for(j=0; j<5; j++)
t[i] ^= statechainv[i+8*j];
#pragma unroll 5
for(j=0;j<5;j++) {
t[i] ^= state->chainv[i+8*j];
}
}
MULT2(t, 0);
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
statechainv[i+8*j] ^= t[i];
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+8*j] ^= t[i];
}
}
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
t[i+8*j] = statechainv[i+8*j];
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
t[i+8*j] = state->chainv[i+8*j];
}
}
#pragma unroll
for(j=0; j<5; j++)
MULT2(statechainv, j);
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
statechainv[8*j+i] ^= t[8*((j+1)%5)+i];
#pragma unroll 5
for(j=0;j<5;j++) {
MULT2(state->chainv, j);
}
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
t[i+8*j] = statechainv[i+8*j];
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[8*j+i] ^= t[8*((j+1)%5)+i];
}
}
#pragma unroll
for(j=0; j<5; j++)
MULT2(statechainv, j);
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
statechainv[8*j+i] ^= t[8*((j+4)%5)+i];
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
t[i+8*j] = state->chainv[i+8*j];
}
}
#pragma unroll 5
for(j=0; j<5; j++) {
#pragma unroll 8
for(i=0; i<8; i++)
statechainv[i+8*j] ^= statebuffer[i];
MULT2(statebuffer, 0);
#pragma unroll 5
for(j=0;j<5;j++) {
MULT2(state->chainv, j);
}
#pragma unroll
for(i=0; i<8; i++)
chainv[i] = statechainv[i];
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[8*j+i] ^= t[8*((j+4)%5)+i];
}
}
#pragma unroll
for(i=0; i<8; i++)
STEP(c_CNS[(2*i)], c_CNS[(2*i)+1]);
#pragma unroll 5
for(j=0;j<5;j++) {
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+8*j] ^= state->buffer[i];
}
MULT2(state->buffer, 0);
}
#pragma unroll
for(i=0; i<8; i++) {
statechainv[i] = chainv[i];
chainv[i] = statechainv[i+8];
#pragma unroll 8
for(i=0;i<8;i++) {
chainv[i] = state->chainv[i];
}
#pragma unroll 8
for(i=0;i<8;i++) {
STEP(c_CNS[(2*i)],c_CNS[(2*i)+1]);
}
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i] = chainv[i];
chainv[i] = state->chainv[i+8];
}
TWEAK(chainv[4],chainv[5],chainv[6],chainv[7],1);
#pragma unroll
for(i=0; i<8; i++)
STEP(c_CNS[(2*i)+16], c_CNS[(2*i)+16+1]);
#pragma unroll 8
for(i=0;i<8;i++) {
STEP(c_CNS[(2*i)+16],c_CNS[(2*i)+16+1]);
}
#pragma unroll
for(i=0; i<8; i++) {
statechainv[i+8] = chainv[i];
chainv[i] = statechainv[i+16];
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+8] = chainv[i];
chainv[i] = state->chainv[i+16];
}
TWEAK(chainv[4],chainv[5],chainv[6],chainv[7],2);
#pragma unroll
for(i=0; i<8; i++)
#pragma unroll 8
for(i=0;i<8;i++) {
STEP(c_CNS[(2*i)+32],c_CNS[(2*i)+32+1]);
}
#pragma unroll
for(i=0; i<8; i++) {
statechainv[i+16] = chainv[i];
chainv[i] = statechainv[i+24];
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+16] = chainv[i];
chainv[i] = state->chainv[i+24];
}
TWEAK(chainv[4],chainv[5],chainv[6],chainv[7],3);
#pragma unroll
for(i=0; i<8; i++)
#pragma unroll 8
for(i=0;i<8;i++) {
STEP(c_CNS[(2*i)+48],c_CNS[(2*i)+48+1]);
}
#pragma unroll
for(i=0; i<8; i++) {
statechainv[i+24] = chainv[i];
chainv[i] = statechainv[i+32];
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+24] = chainv[i];
chainv[i] = state->chainv[i+32];
}
TWEAK(chainv[4],chainv[5],chainv[6],chainv[7],4);
#pragma unroll
for(i=0; i<8; i++)
#pragma unroll 8
for(i=0;i<8;i++) {
STEP(c_CNS[(2*i)+64],c_CNS[(2*i)+64+1]);
}
#pragma unroll 8
for(i=0; i<8; i++)
statechainv[i+32] = chainv[i];
#pragma unroll 8
for(i=0;i<8;i++) {
state->chainv[i+32] = chainv[i];
}
}
static void rnd512_cpu(uint32_t *statebuffer, uint32_t *statechainv)
{
int i, j;
uint32_t t[40];
uint32_t chainv[8];
uint32_t tmp;
for (i = 0; i<8; i++) {
t[i] = statechainv[i];
for (j = 1; j<5; j++)
t[i] ^= statechainv[i + 8 * j];
}
MULT2(t, 0);
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
statechainv[i + 8 * j] ^= t[i];
}
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
t[i + 8 * j] = statechainv[i + 8 * j];
}
for (j = 0; j<5; j++)
MULT2(statechainv, j);
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
statechainv[8 * j + i] ^= t[8 * ((j + 1) % 5) + i];
}
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
t[i + 8 * j] = statechainv[i + 8 * j];
}
for (j = 0; j<5; j++)
MULT2(statechainv, j);
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
statechainv[8 * j + i] ^= t[8 * ((j + 4) % 5) + i];
}
for (j = 0; j<5; j++) {
for (i = 0; i<8; i++)
statechainv[i + 8 * j] ^= statebuffer[i];
MULT2(statebuffer, 0);
}
for (i = 0; i<8; i++)
chainv[i] = statechainv[i];
for (i = 0; i<8; i++)
STEP(h_CNS[(2 * i)], h_CNS[(2 * i) + 1]);
for (i = 0; i<8; i++) {
statechainv[i] = chainv[i];
chainv[i] = statechainv[i + 8];
}
TWEAK(chainv[4], chainv[5], chainv[6], chainv[7], 1);
for (i = 0; i<8; i++)
STEP(h_CNS[(2 * i) + 16], h_CNS[(2 * i) + 16 + 1]);
for (i = 0; i<8; i++) {
statechainv[i + 8] = chainv[i];
chainv[i] = statechainv[i + 16];
}
TWEAK(chainv[4], chainv[5], chainv[6], chainv[7], 2);
for (i = 0; i<8; i++)
STEP(h_CNS[(2 * i) + 32], h_CNS[(2 * i) + 32 + 1]);
for (i = 0; i<8; i++) {
statechainv[i + 16] = chainv[i];
chainv[i] = statechainv[i + 24];
}
TWEAK(chainv[4], chainv[5], chainv[6], chainv[7], 3);
for (i = 0; i<8; i++)
STEP(h_CNS[(2 * i) + 48], h_CNS[(2 * i) + 48 + 1]);
for (i = 0; i<8; i++) {
statechainv[i + 24] = chainv[i];
chainv[i] = statechainv[i + 32];
}
TWEAK(chainv[4], chainv[5], chainv[6], chainv[7], 4);
for (i = 0; i<8; i++)
STEP(h_CNS[(2 * i) + 64], h_CNS[(2 * i) + 64 + 1]);
for (i = 0; i<8; i++)
statechainv[i + 32] = chainv[i];
}
/***************************************************/
__device__ __forceinline__
void Update512(uint32_t* statebuffer, uint32_t *statechainv, const uint32_t *const __restrict__ data)
void Update512(hashState *state, const BitSequence *data)
{
#pragma unroll
for (int i = 0; i<8; i++)
statebuffer[i] = cuda_swab32((data[i]));
rnd512(statebuffer, statechainv);
#pragma unroll 8
for(int i=0;i<8;i++) state->buffer[i] = BYTES_SWAP32(((uint32_t*)data)[i]);
rnd512(state);
#pragma unroll
for(int i=0; i<8; i++)
statebuffer[i] = cuda_swab32((data[i+8]));
rnd512(statebuffer, statechainv);
#pragma unroll
for(int i=0; i<4; i++)
statebuffer[i] = cuda_swab32((data[i+16]));
#pragma unroll 8
for(int i=0;i<8;i++) state->buffer[i] = BYTES_SWAP32(((uint32_t*)(data+32))[i]);
rnd512(state);
#pragma unroll 4
for(int i=0;i<4;i++) state->buffer[i] = BYTES_SWAP32(((uint32_t*)(data+64))[i]);
}
/***************************************************/
__device__ __forceinline__
void finalization512(uint32_t* statebuffer, uint32_t *statechainv, uint32_t *b)
void finalization512(hashState *state, uint32_t *b)
{
int i,j;
statebuffer[4] = 0x80000000U;
#pragma unroll 3
for(int i=5; i<8; i++)
statebuffer[i] = 0;
rnd512(statebuffer, statechainv);
state->buffer[4] = 0x80000000;
#pragma unroll 3
for(int i=5;i<8;i++) state->buffer[i] = 0;
rnd512(state);
/*---- blank round with m=0 ----*/
#pragma unroll
for(i=0; i<8; i++)
statebuffer[i] =0;
rnd512(statebuffer, statechainv);
#pragma unroll 8
for(i=0;i<8;i++) state->buffer[i] =0;
rnd512(state);
#pragma unroll
for(i=0; i<8; i++) {
#pragma unroll 8
for(i=0;i<8;i++) {
b[i] = 0;
#pragma unroll 5
for(j=0; j<5; j++)
b[i] ^= statechainv[i+8*j];
b[i] = cuda_swab32((b[i]));
#pragma unroll 5
for(j=0;j<5;j++) {
b[i] ^= state->chainv[i+8*j];
}
b[i] = BYTES_SWAP32((b[i]));
}
#pragma unroll
for(i=0; i<8; i++)
statebuffer[i]=0;
rnd512(statebuffer, statechainv);
#pragma unroll 8
for(i=0;i<8;i++) state->buffer[i]=0;
rnd512(state);
#pragma unroll
for(i=0; i<8; i++)
{
#pragma unroll 8
for(i=0;i<8;i++) {
b[8+i] = 0;
#pragma unroll 5
for(j=0; j<5; j++)
b[8+i] ^= statechainv[i+8*j];
b[8+i] = cuda_swab32((b[8+i]));
#pragma unroll 5
for(j=0;j<5;j++) {
b[8+i] ^= state->chainv[i+8*j];
}
b[8+i] = BYTES_SWAP32((b[8+i]));
}
}
/***************************************************/
// Die Hash-Funktion
__global__
void qubit_luffa512_gpu_hash_80(uint32_t threads, uint32_t startNounce, uint32_t *outputHash)
void qubit_luffa512_gpu_hash_80(uint32_t threads, uint32_t startNounce, void *outputHash)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
@@ -423,37 +359,118 @@ void qubit_luffa512_gpu_hash_80(uint32_t threads, uint32_t startNounce, uint32_t
uint32_t buf32[32];
} buff;
#pragma unroll 8
for (int i=8; i < 16; i++)
buff.buf64[i] = c_PaddedMessage80[i];
#pragma unroll 16
for (int i=0; i < 16; ++i) buff.buf64[i] = c_PaddedMessage80[i];
// die Nounce durch die thread-spezifische ersetzen
buff.buf64[9] = REPLACE_HIDWORD(buff.buf64[9], cuda_swab32(nounce));
uint32_t statebuffer[8], statechainv[40];
#pragma unroll
for (int i = 0; i<4; i++)
statebuffer[i] = cuda_swab32(buff.buf32[i + 16]);
hashState state;
#pragma unroll 40
for(int i=0;i<40;i++) state.chainv[i] = c_IV[i];
#pragma unroll 8
for(int i=0;i<8;i++) state.buffer[i] = 0;
Update512(&state, (BitSequence*)buff.buf32);
uint32_t *outHash = (uint32_t *)outputHash + 16 * thread;
finalization512(&state, (uint32_t*)outHash);
}
}
#pragma unroll 4
for (int i = 4; i<8; i++)
statebuffer[i] = statebufferpre[i];
__global__
void qubit_luffa512_gpu_finalhash_80(uint32_t threads, uint32_t startNounce, void *outputHash, uint32_t *resNounce)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint32_t nounce = startNounce + thread;
union {
uint64_t buf64[16];
uint32_t buf32[32];
} buff;
uint32_t Hash[16];
#pragma unroll
for (int i = 0; i<40; i++)
statechainv[i] = statechainvpre[i];
#pragma unroll 16
for (int i=0; i < 16; ++i) buff.buf64[i] = c_PaddedMessage80[i];
uint32_t *outHash = &outputHash[thread * 16];
finalization512(statebuffer, statechainv, outHash);
// Tested nonce
buff.buf64[9] = REPLACE_HIDWORD(buff.buf64[9], cuda_swab32(nounce));
hashState state;
#pragma unroll 40
for(int i=0;i<40;i++) state.chainv[i] = c_IV[i];
#pragma unroll 8
for(int i=0;i<8;i++) state.buffer[i] = 0;
Update512(&state, (BitSequence*)buff.buf32);
finalization512(&state, Hash);
/* dont ask me why not a simple if (Hash[i] > c_Target[i]) return;
* we lose 20% in perfs without the position test */
int position = -1;
#pragma unroll 8
for (int i = 7; i >= 0; i--) {
if (Hash[i] > c_Target[i]) {
if (position < i) {
return;
}
}
if (Hash[i] < c_Target[i]) {
if (position < i) {
position = i;
//break; /* impact perfs, unroll ? */
}
}
}
#if NBN == 1
if (resNounce[0] > nounce) {
resNounce[0] = nounce;
}
#else
/* keep the smallest nounce, + extra one if found */
if (resNounce[0] > nounce) {
resNounce[1] = resNounce[0];
resNounce[0] = nounce;
} else {
resNounce[1] = nounce;
}
#endif
}
}
__host__
void qubit_luffa512_cpu_init(int thr_id, uint32_t threads)
{
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_IV, h_IV, sizeof(h_IV), 0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_CNS, h_CNS, sizeof(h_CNS), 0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_IV, h2_IV, sizeof(h2_IV), 0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_CNS, h2_CNS, sizeof(h2_CNS), 0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMalloc(&d_resNounce[thr_id], NBN * sizeof(uint32_t)));
CUDA_SAFE_CALL(cudaMallocHost(&h_resNounce[thr_id], NBN * sizeof(uint32_t)));
}
__host__
uint32_t qubit_luffa512_cpu_finalhash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_outputHash,int order)
{
uint32_t result = UINT32_MAX;
cudaMemset(d_resNounce[thr_id], 0xff, NBN * sizeof(uint32_t));
const uint32_t threadsperblock = 256;
dim3 grid((threads + threadsperblock-1)/threadsperblock);
dim3 block(threadsperblock);
size_t shared_size = 0;
qubit_luffa512_gpu_finalhash_80 <<<grid, block, shared_size>>> (threads, startNounce, d_outputHash, d_resNounce[thr_id]);
cudaThreadSynchronize();
if (cudaSuccess == cudaMemcpy(h_resNounce[thr_id], d_resNounce[thr_id], NBN * sizeof(uint32_t), cudaMemcpyDeviceToHost)) {
//cudaThreadSynchronize();
result = h_resNounce[thr_id][0];
#if NBN > 1
extra_results[0] = h_resNounce[thr_id][1];
#endif
}
return result;
}
__host__
@@ -468,37 +485,6 @@ void qubit_luffa512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNoun
qubit_luffa512_gpu_hash_80 <<<grid, block, shared_size>>> (threads, startNounce, d_outputHash);
}
__host__
void qubit_cpu_precalc(uint32_t* message)
{
uint32_t statebuffer[8];
uint32_t statechainv[40] =
{
0x6d251e69, 0x44b051e0, 0x4eaa6fb4, 0xdbf78465,
0x6e292011, 0x90152df4, 0xee058139, 0xdef610bb,
0xc3b44b95, 0xd9d2f256, 0x70eee9a0, 0xde099fa3,
0x5d9b0557, 0x8fc944b3, 0xcf1ccf0e, 0x746cd581,
0xf7efc89d, 0x5dba5781, 0x04016ce5, 0xad659c05,
0x0306194f, 0x666d1836, 0x24aa230a, 0x8b264ae7,
0x858075d5, 0x36d79cce, 0xe571f7d7, 0x204b1f67,
0x35870c6a, 0x57e9e923, 0x14bcb808, 0x7cde72ce,
0x6c68e9be, 0x5ec41e22, 0xc825b7c7, 0xaffb4363,
0xf5df3999, 0x0fc688f1, 0xb07224cc, 0x03e86cea
};
for (int i = 0; i<8; i++)
statebuffer[i] = cuda_swab32(message[i]);
rnd512_cpu(statebuffer, statechainv);
for (int i = 0; i<8; i++)
statebuffer[i] = cuda_swab32(message[i+8]);
rnd512_cpu(statebuffer, statechainv);
cudaMemcpyToSymbol(statebufferpre, statebuffer, sizeof(statebuffer), 0, cudaMemcpyHostToDevice);
cudaMemcpyToSymbol(statechainvpre, statechainv, sizeof(statechainv), 0, cudaMemcpyHostToDevice);
}
__host__
void qubit_luffa512_cpu_setBlock_80(void *pdata)
{
@@ -511,6 +497,21 @@ void qubit_luffa512_cpu_setBlock_80(void *pdata)
PaddedMessage[126] = 0x02;
PaddedMessage[127] = 0x80;
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_PaddedMessage80, PaddedMessage, sizeof(PaddedMessage), 0, cudaMemcpyHostToDevice));
qubit_cpu_precalc((uint32_t*) PaddedMessage);
CUDA_SAFE_CALL(cudaMemcpyToSymbol( c_PaddedMessage80, PaddedMessage, 16*sizeof(uint64_t), 0, cudaMemcpyHostToDevice));
}
__host__
void qubit_luffa512_cpufinal_setBlock_80(void *pdata, const void *ptarget)
{
unsigned char PaddedMessage[128];
memcpy(PaddedMessage, pdata, 80);
memset(PaddedMessage+80, 0, 48);
PaddedMessage[80] = 0x80;
PaddedMessage[111] = 1;
PaddedMessage[126] = 0x02;
PaddedMessage[127] = 0x80;
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_Target, ptarget, 8*sizeof(uint32_t), 0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_PaddedMessage80, PaddedMessage, 16*sizeof(uint64_t), 0, cudaMemcpyHostToDevice));
}
+4 -4
View File
@@ -60,8 +60,8 @@ IDI_ICON1 ICON "ccminer.ico"
//
VS_VERSION_INFO VERSIONINFO
FILEVERSION 1,6,5,0
PRODUCTVERSION 1,6,5,0
FILEVERSION 1,6,0,0
PRODUCTVERSION 1,6,0,0
FILEFLAGSMASK 0x3fL
#ifdef _DEBUG
FILEFLAGS 0x21L
@@ -76,10 +76,10 @@ BEGIN
BEGIN
BLOCK "040904e4"
BEGIN
VALUE "FileVersion", "1.6.5"
VALUE "FileVersion", "1.6"
VALUE "LegalCopyright", "Copyright (C) 2015"
VALUE "ProductName", "ccminer"
VALUE "ProductVersion", "1.6.5"
VALUE "ProductVersion", "1.6"
END
END
BLOCK "VarFileInfo"
+1
View File
@@ -23,6 +23,7 @@ extern char *device_name[MAX_GPUS];
extern bool opt_autotune;
extern int opt_nfactor;
extern char *jane_params;
extern bool abort_flag;
extern int parallel;
extern void get_currentalgo(char* buf, int sz);
-860
View File
@@ -1,860 +0,0 @@
// Copyright (c) 2009-2010 Satoshi Nakamoto
// Copyright (c) 2009-2014 The Bitcoin Core developers
// Distributed under the MIT software license, see the accompanying
// file COPYING or http://www.opensource.org/licenses/mit-license.php.
#ifndef BITCOIN_SERIALIZE_H
#define BITCOIN_SERIALIZE_H
//#include "compat/endian.h"
#include <algorithm>
#include <assert.h>
#include <ios>
#include <limits>
#include <map>
#include <set>
#include <stdint.h>
#include <string>
#include <string.h>
#include <utility>
#include <vector>
class CScript;
static const unsigned int MAX_SIZE = 0x02000000;
/**
* Used to bypass the rule against non-const reference to temporary
* where it makes sense with wrappers such as CFlatData or CTxDB
*/
template<typename T>
inline T& REF(const T& val)
{
return const_cast<T&>(val);
}
/**
* Used to acquire a non-const pointer "this" to generate bodies
* of const serialization operations from a template
*/
template<typename T>
inline T* NCONST_PTR(const T* val)
{
return const_cast<T*>(val);
}
/**
* Get begin pointer of vector (non-const version).
* @note These functions avoid the undefined case of indexing into an empty
* vector, as well as that of indexing after the end of the vector.
*/
template <class T, class TAl>
inline T* begin_ptr(std::vector<T,TAl>& v)
{
return v.empty() ? NULL : &v[0];
}
/** Get begin pointer of vector (const version) */
template <class T, class TAl>
inline const T* begin_ptr(const std::vector<T,TAl>& v)
{
return v.empty() ? NULL : &v[0];
}
/** Get end pointer of vector (non-const version) */
template <class T, class TAl>
inline T* end_ptr(std::vector<T,TAl>& v)
{
return v.empty() ? NULL : (&v[0] + v.size());
}
/** Get end pointer of vector (const version) */
template <class T, class TAl>
inline const T* end_ptr(const std::vector<T,TAl>& v)
{
return v.empty() ? NULL : (&v[0] + v.size());
}
/*
* Lowest-level serialization and conversion.
* @note Sizes of these types are verified in the tests
*/
template<typename Stream> inline void ser_writedata8(Stream &s, uint8_t obj)
{
s.write((char*)&obj, 1);
}
template<typename Stream> inline void ser_writedata16(Stream &s, uint16_t obj)
{
obj = htole16(obj);
s.write((char*)&obj, 2);
}
template<typename Stream> inline void ser_writedata32(Stream &s, uint32_t obj)
{
obj = htole32(obj);
s.write((char*)&obj, 4);
}
template<typename Stream> inline void ser_writedata64(Stream &s, uint64_t obj)
{
obj = htole64(obj);
s.write((char*)&obj, 8);
}
template<typename Stream> inline uint8_t ser_readdata8(Stream &s)
{
uint8_t obj;
s.read((char*)&obj, 1);
return obj;
}
template<typename Stream> inline uint16_t ser_readdata16(Stream &s)
{
uint16_t obj;
s.read((char*)&obj, 2);
return le16toh(obj);
}
template<typename Stream> inline uint32_t ser_readdata32(Stream &s)
{
uint32_t obj;
s.read((char*)&obj, 4);
return le32toh(obj);
}
template<typename Stream> inline uint64_t ser_readdata64(Stream &s)
{
uint64_t obj;
s.read((char*)&obj, 8);
return le64toh(obj);
}
inline uint64_t ser_double_to_uint64(double x)
{
union { double x; uint64_t y; } tmp;
tmp.x = x;
return tmp.y;
}
inline uint32_t ser_float_to_uint32(float x)
{
union { float x; uint32_t y; } tmp;
tmp.x = x;
return tmp.y;
}
inline double ser_uint64_to_double(uint64_t y)
{
union { double x; uint64_t y; } tmp;
tmp.y = y;
return tmp.x;
}
inline float ser_uint32_to_float(uint32_t y)
{
union { float x; uint32_t y; } tmp;
tmp.y = y;
return tmp.x;
}
/////////////////////////////////////////////////////////////////
//
// Templates for serializing to anything that looks like a stream,
// i.e. anything that supports .read(char*, size_t) and .write(char*, size_t)
//
enum
{
// primary actions
SER_NETWORK = (1 << 0),
SER_DISK = (1 << 1),
SER_GETHASH = (1 << 2),
};
#define READWRITE(obj) (::SerReadWrite(s, (obj), nType, nVersion, ser_action))
/**
* Implement three methods for serializable objects. These are actually wrappers over
* "SerializationOp" template, which implements the body of each class' serialization
* code. Adding "ADD_SERIALIZE_METHODS" in the body of the class causes these wrappers to be
* added as members.
*/
#define ADD_SERIALIZE_METHODS \
size_t GetSerializeSize(int nType, int nVersion) const { \
CSizeComputer s(nType, nVersion); \
NCONST_PTR(this)->SerializationOp(s, CSerActionSerialize(), nType, nVersion);\
return s.size(); \
} \
template<typename Stream> \
void Serialize(Stream& s, int nType, int nVersion) const { \
NCONST_PTR(this)->SerializationOp(s, CSerActionSerialize(), nType, nVersion);\
} \
template<typename Stream> \
void Unserialize(Stream& s, int nType, int nVersion) { \
SerializationOp(s, CSerActionUnserialize(), nType, nVersion); \
}
/*
* Basic Types
*/
inline unsigned int GetSerializeSize(char a, int, int=0) { return 1; }
inline unsigned int GetSerializeSize(int8_t a, int, int=0) { return 1; }
inline unsigned int GetSerializeSize(uint8_t a, int, int=0) { return 1; }
inline unsigned int GetSerializeSize(int16_t a, int, int=0) { return 2; }
inline unsigned int GetSerializeSize(uint16_t a, int, int=0) { return 2; }
inline unsigned int GetSerializeSize(int32_t a, int, int=0) { return 4; }
inline unsigned int GetSerializeSize(uint32_t a, int, int=0) { return 4; }
inline unsigned int GetSerializeSize(int64_t a, int, int=0) { return 8; }
inline unsigned int GetSerializeSize(uint64_t a, int, int=0) { return 8; }
inline unsigned int GetSerializeSize(float a, int, int=0) { return 4; }
inline unsigned int GetSerializeSize(double a, int, int=0) { return 8; }
template<typename Stream> inline void Serialize(Stream& s, char a, int, int=0) { ser_writedata8(s, a); } // TODO Get rid of bare char
template<typename Stream> inline void Serialize(Stream& s, int8_t a, int, int=0) { ser_writedata8(s, a); }
template<typename Stream> inline void Serialize(Stream& s, uint8_t a, int, int=0) { ser_writedata8(s, a); }
template<typename Stream> inline void Serialize(Stream& s, int16_t a, int, int=0) { ser_writedata16(s, a); }
template<typename Stream> inline void Serialize(Stream& s, uint16_t a, int, int=0) { ser_writedata16(s, a); }
template<typename Stream> inline void Serialize(Stream& s, int32_t a, int, int=0) { ser_writedata32(s, a); }
template<typename Stream> inline void Serialize(Stream& s, uint32_t a, int, int=0) { ser_writedata32(s, a); }
template<typename Stream> inline void Serialize(Stream& s, int64_t a, int, int=0) { ser_writedata64(s, a); }
template<typename Stream> inline void Serialize(Stream& s, uint64_t a, int, int=0) { ser_writedata64(s, a); }
template<typename Stream> inline void Serialize(Stream& s, float a, int, int=0) { ser_writedata32(s, ser_float_to_uint32(a)); }
template<typename Stream> inline void Serialize(Stream& s, double a, int, int=0) { ser_writedata64(s, ser_double_to_uint64(a)); }
template<typename Stream> inline void Unserialize(Stream& s, char& a, int, int=0) { a = ser_readdata8(s); } // TODO Get rid of bare char
template<typename Stream> inline void Unserialize(Stream& s, int8_t& a, int, int=0) { a = ser_readdata8(s); }
template<typename Stream> inline void Unserialize(Stream& s, uint8_t& a, int, int=0) { a = ser_readdata8(s); }
template<typename Stream> inline void Unserialize(Stream& s, int16_t& a, int, int=0) { a = ser_readdata16(s); }
template<typename Stream> inline void Unserialize(Stream& s, uint16_t& a, int, int=0) { a = ser_readdata16(s); }
template<typename Stream> inline void Unserialize(Stream& s, int32_t& a, int, int=0) { a = ser_readdata32(s); }
template<typename Stream> inline void Unserialize(Stream& s, uint32_t& a, int, int=0) { a = ser_readdata32(s); }
template<typename Stream> inline void Unserialize(Stream& s, int64_t& a, int, int=0) { a = ser_readdata64(s); }
template<typename Stream> inline void Unserialize(Stream& s, uint64_t& a, int, int=0) { a = ser_readdata64(s); }
template<typename Stream> inline void Unserialize(Stream& s, float& a, int, int=0) { a = ser_uint32_to_float(ser_readdata32(s)); }
template<typename Stream> inline void Unserialize(Stream& s, double& a, int, int=0) { a = ser_uint64_to_double(ser_readdata64(s)); }
inline unsigned int GetSerializeSize(bool a, int, int=0) { return sizeof(char); }
template<typename Stream> inline void Serialize(Stream& s, bool a, int, int=0) { char f=a; ser_writedata8(s, f); }
template<typename Stream> inline void Unserialize(Stream& s, bool& a, int, int=0) { char f=ser_readdata8(s); a=f; }
/**
* Compact Size
* size < 253 -- 1 byte
* size <= USHRT_MAX -- 3 bytes (253 + 2 bytes)
* size <= UINT_MAX -- 5 bytes (254 + 4 bytes)
* size > UINT_MAX -- 9 bytes (255 + 8 bytes)
*/
inline unsigned int GetSizeOfCompactSize(uint64_t nSize)
{
if (nSize < 253) return sizeof(unsigned char);
else if (nSize <= std::numeric_limits<unsigned short>::max()) return sizeof(unsigned char) + sizeof(unsigned short);
else if (nSize <= std::numeric_limits<unsigned int>::max()) return sizeof(unsigned char) + sizeof(unsigned int);
else return sizeof(unsigned char) + sizeof(uint64_t);
}
template<typename Stream>
void WriteCompactSize(Stream& os, uint64_t nSize)
{
if (nSize < 253)
{
ser_writedata8(os, nSize);
}
else if (nSize <= std::numeric_limits<unsigned short>::max())
{
ser_writedata8(os, 253);
ser_writedata16(os, nSize);
}
else if (nSize <= std::numeric_limits<unsigned int>::max())
{
ser_writedata8(os, 254);
ser_writedata32(os, nSize);
}
else
{
ser_writedata8(os, 255);
ser_writedata64(os, nSize);
}
return;
}
template<typename Stream>
uint64_t ReadCompactSize(Stream& is)
{
uint8_t chSize = ser_readdata8(is);
uint64_t nSizeRet = 0;
if (chSize < 253)
{
nSizeRet = chSize;
}
else if (chSize == 253)
{
nSizeRet = ser_readdata16(is);
if (nSizeRet < 253)
throw std::ios_base::failure("non-canonical ReadCompactSize()");
}
else if (chSize == 254)
{
nSizeRet = ser_readdata32(is);
if (nSizeRet < 0x10000u)
throw std::ios_base::failure("non-canonical ReadCompactSize()");
}
else
{
nSizeRet = ser_readdata64(is);
if (nSizeRet < 0x100000000ULL)
throw std::ios_base::failure("non-canonical ReadCompactSize()");
}
if (nSizeRet > (uint64_t)MAX_SIZE)
throw std::ios_base::failure("ReadCompactSize(): size too large");
return nSizeRet;
}
/**
* Variable-length integers: bytes are a MSB base-128 encoding of the number.
* The high bit in each byte signifies whether another digit follows. To make
* sure the encoding is one-to-one, one is subtracted from all but the last digit.
* Thus, the byte sequence a[] with length len, where all but the last byte
* has bit 128 set, encodes the number:
*
* (a[len-1] & 0x7F) + sum(i=1..len-1, 128^i*((a[len-i-1] & 0x7F)+1))
*
* Properties:
* * Very small (0-127: 1 byte, 128-16511: 2 bytes, 16512-2113663: 3 bytes)
* * Every integer has exactly one encoding
* * Encoding does not depend on size of original integer type
* * No redundancy: every (infinite) byte sequence corresponds to a list
* of encoded integers.
*
* 0: [0x00] 256: [0x81 0x00]
* 1: [0x01] 16383: [0xFE 0x7F]
* 127: [0x7F] 16384: [0xFF 0x00]
* 128: [0x80 0x00] 16511: [0x80 0xFF 0x7F]
* 255: [0x80 0x7F] 65535: [0x82 0xFD 0x7F]
* 2^32: [0x8E 0xFE 0xFE 0xFF 0x00]
*/
template<typename I>
inline unsigned int GetSizeOfVarInt(I n)
{
int nRet = 0;
while(true) {
nRet++;
if (n <= 0x7F)
break;
n = (n >> 7) - 1;
}
return nRet;
}
template<typename Stream, typename I>
void WriteVarInt(Stream& os, I n)
{
unsigned char tmp[(sizeof(n)*8+6)/7];
int len=0;
while(true) {
tmp[len] = (n & 0x7F) | (len ? 0x80 : 0x00);
if (n <= 0x7F)
break;
n = (n >> 7) - 1;
len++;
}
do {
ser_writedata8(os, tmp[len]);
} while(len--);
}
template<typename Stream, typename I>
I ReadVarInt(Stream& is)
{
I n = 0;
while(true) {
unsigned char chData = ser_readdata8(is);
n = (n << 7) | (chData & 0x7F);
if (chData & 0x80)
n++;
else
return n;
}
}
#define FLATDATA(obj) REF(CFlatData((char*)&(obj), (char*)&(obj) + sizeof(obj)))
#define VARINT(obj) REF(WrapVarInt(REF(obj)))
#define LIMITED_STRING(obj,n) REF(LimitedString< n >(REF(obj)))
/**
* Wrapper for serializing arrays and POD.
*/
class CFlatData
{
protected:
char* pbegin;
char* pend;
public:
CFlatData(void* pbeginIn, void* pendIn) : pbegin((char*)pbeginIn), pend((char*)pendIn) { }
template <class T, class TAl>
explicit CFlatData(std::vector<T,TAl> &v)
{
pbegin = (char*)begin_ptr(v);
pend = (char*)end_ptr(v);
}
char* begin() { return pbegin; }
const char* begin() const { return pbegin; }
char* end() { return pend; }
const char* end() const { return pend; }
unsigned int GetSerializeSize(int, int=0) const
{
return pend - pbegin;
}
template<typename Stream>
void Serialize(Stream& s, int, int=0) const
{
s.write(pbegin, pend - pbegin);
}
template<typename Stream>
void Unserialize(Stream& s, int, int=0)
{
s.read(pbegin, pend - pbegin);
}
};
template<typename I>
class CVarInt
{
protected:
I &n;
public:
CVarInt(I& nIn) : n(nIn) { }
unsigned int GetSerializeSize(int, int) const {
return GetSizeOfVarInt<I>(n);
}
template<typename Stream>
void Serialize(Stream &s, int, int) const {
WriteVarInt<Stream,I>(s, n);
}
template<typename Stream>
void Unserialize(Stream& s, int, int) {
n = ReadVarInt<Stream,I>(s);
}
};
template<size_t Limit>
class LimitedString
{
protected:
std::string& string;
public:
LimitedString(std::string& string) : string(string) {}
template<typename Stream>
void Unserialize(Stream& s, int, int=0)
{
size_t size = ReadCompactSize(s);
if (size > Limit) {
throw std::ios_base::failure("String length limit exceeded");
}
string.resize(size);
if (size != 0)
s.read((char*)&string[0], size);
}
template<typename Stream>
void Serialize(Stream& s, int, int=0) const
{
WriteCompactSize(s, string.size());
if (!string.empty())
s.write((char*)&string[0], string.size());
}
unsigned int GetSerializeSize(int, int=0) const
{
return GetSizeOfCompactSize(string.size()) + string.size();
}
};
template<typename I>
CVarInt<I> WrapVarInt(I& n) { return CVarInt<I>(n); }
/**
* Forward declarations
*/
/**
* string
*/
template<typename C> unsigned int GetSerializeSize(const std::basic_string<C>& str, int, int=0);
template<typename Stream, typename C> void Serialize(Stream& os, const std::basic_string<C>& str, int, int=0);
template<typename Stream, typename C> void Unserialize(Stream& is, std::basic_string<C>& str, int, int=0);
/**
* vector
* vectors of unsigned char are a special case and are intended to be serialized as a single opaque blob.
*/
template<typename T, typename A> unsigned int GetSerializeSize_impl(const std::vector<T, A>& v, int nType, int nVersion, const unsigned char&);
template<typename T, typename A, typename V> unsigned int GetSerializeSize_impl(const std::vector<T, A>& v, int nType, int nVersion, const V&);
template<typename T, typename A> inline unsigned int GetSerializeSize(const std::vector<T, A>& v, int nType, int nVersion);
template<typename Stream, typename T, typename A> void Serialize_impl(Stream& os, const std::vector<T, A>& v, int nType, int nVersion, const unsigned char&);
template<typename Stream, typename T, typename A, typename V> void Serialize_impl(Stream& os, const std::vector<T, A>& v, int nType, int nVersion, const V&);
template<typename Stream, typename T, typename A> inline void Serialize(Stream& os, const std::vector<T, A>& v, int nType, int nVersion);
template<typename Stream, typename T, typename A> void Unserialize_impl(Stream& is, std::vector<T, A>& v, int nType, int nVersion, const unsigned char&);
template<typename Stream, typename T, typename A, typename V> void Unserialize_impl(Stream& is, std::vector<T, A>& v, int nType, int nVersion, const V&);
template<typename Stream, typename T, typename A> inline void Unserialize(Stream& is, std::vector<T, A>& v, int nType, int nVersion);
/**
* others derived from vector
*/
extern inline unsigned int GetSerializeSize(const CScript& v, int nType, int nVersion);
template<typename Stream> void Serialize(Stream& os, const CScript& v, int nType, int nVersion);
template<typename Stream> void Unserialize(Stream& is, CScript& v, int nType, int nVersion);
/**
* pair
*/
template<typename K, typename T> unsigned int GetSerializeSize(const std::pair<K, T>& item, int nType, int nVersion);
template<typename Stream, typename K, typename T> void Serialize(Stream& os, const std::pair<K, T>& item, int nType, int nVersion);
template<typename Stream, typename K, typename T> void Unserialize(Stream& is, std::pair<K, T>& item, int nType, int nVersion);
/**
* map
*/
template<typename K, typename T, typename Pred, typename A> unsigned int GetSerializeSize(const std::map<K, T, Pred, A>& m, int nType, int nVersion);
template<typename Stream, typename K, typename T, typename Pred, typename A> void Serialize(Stream& os, const std::map<K, T, Pred, A>& m, int nType, int nVersion);
template<typename Stream, typename K, typename T, typename Pred, typename A> void Unserialize(Stream& is, std::map<K, T, Pred, A>& m, int nType, int nVersion);
/**
* set
*/
template<typename K, typename Pred, typename A> unsigned int GetSerializeSize(const std::set<K, Pred, A>& m, int nType, int nVersion);
template<typename Stream, typename K, typename Pred, typename A> void Serialize(Stream& os, const std::set<K, Pred, A>& m, int nType, int nVersion);
template<typename Stream, typename K, typename Pred, typename A> void Unserialize(Stream& is, std::set<K, Pred, A>& m, int nType, int nVersion);
/**
* If none of the specialized versions above matched, default to calling member function.
* "int nType" is changed to "long nType" to keep from getting an ambiguous overload error.
* The compiler will only cast int to long if none of the other templates matched.
* Thanks to Boost serialization for this idea.
*/
template<typename T>
inline unsigned int GetSerializeSize(const T& a, long nType, int nVersion)
{
return a.GetSerializeSize((int)nType, nVersion);
}
template<typename Stream, typename T>
inline void Serialize(Stream& os, const T& a, long nType, int nVersion)
{
a.Serialize(os, (int)nType, nVersion);
}
template<typename Stream, typename T>
inline void Unserialize(Stream& is, T& a, long nType, int nVersion)
{
a.Unserialize(is, (int)nType, nVersion);
}
/**
* string
*/
template<typename C>
unsigned int GetSerializeSize(const std::basic_string<C>& str, int, int)
{
return GetSizeOfCompactSize(str.size()) + str.size() * sizeof(str[0]);
}
template<typename Stream, typename C>
void Serialize(Stream& os, const std::basic_string<C>& str, int, int)
{
WriteCompactSize(os, str.size());
if (!str.empty())
os.write((char*)&str[0], str.size() * sizeof(str[0]));
}
template<typename Stream, typename C>
void Unserialize(Stream& is, std::basic_string<C>& str, int, int)
{
unsigned int nSize = ReadCompactSize(is);
str.resize(nSize);
if (nSize != 0)
is.read((char*)&str[0], nSize * sizeof(str[0]));
}
/**
* vector
*/
template<typename T, typename A>
unsigned int GetSerializeSize_impl(const std::vector<T, A>& v, int nType, int nVersion, const unsigned char&)
{
return (GetSizeOfCompactSize(v.size()) + v.size() * sizeof(T));
}
template<typename T, typename A, typename V>
unsigned int GetSerializeSize_impl(const std::vector<T, A>& v, int nType, int nVersion, const V&)
{
unsigned int nSize = GetSizeOfCompactSize(v.size());
for (typename std::vector<T, A>::const_iterator vi = v.begin(); vi != v.end(); ++vi)
nSize += GetSerializeSize((*vi), nType, nVersion);
return nSize;
}
template<typename T, typename A>
inline unsigned int GetSerializeSize(const std::vector<T, A>& v, int nType, int nVersion)
{
return GetSerializeSize_impl(v, nType, nVersion, T());
}
template<typename Stream, typename T, typename A>
void Serialize_impl(Stream& os, const std::vector<T, A>& v, int nType, int nVersion, const unsigned char&)
{
WriteCompactSize(os, v.size());
if (!v.empty())
os.write((char*)&v[0], v.size() * sizeof(T));
}
template<typename Stream, typename T, typename A, typename V>
void Serialize_impl(Stream& os, const std::vector<T, A>& v, int nType, int nVersion, const V&)
{
WriteCompactSize(os, v.size());
for (typename std::vector<T, A>::const_iterator vi = v.begin(); vi != v.end(); ++vi)
::Serialize(os, (*vi), nType, nVersion);
}
template<typename Stream, typename T, typename A>
inline void Serialize(Stream& os, const std::vector<T, A>& v, int nType, int nVersion)
{
Serialize_impl(os, v, nType, nVersion, T());
}
template<typename Stream, typename T, typename A>
void Unserialize_impl(Stream& is, std::vector<T, A>& v, int nType, int nVersion, const unsigned char&)
{
// Limit size per read so bogus size value won't cause out of memory
v.clear();
unsigned int nSize = ReadCompactSize(is);
unsigned int i = 0;
while (i < nSize)
{
unsigned int blk = std::min(nSize - i, (unsigned int)(1 + 4999999 / sizeof(T)));
v.resize(i + blk);
is.read((char*)&v[i], blk * sizeof(T));
i += blk;
}
}
template<typename Stream, typename T, typename A, typename V>
void Unserialize_impl(Stream& is, std::vector<T, A>& v, int nType, int nVersion, const V&)
{
v.clear();
unsigned int nSize = ReadCompactSize(is);
unsigned int i = 0;
unsigned int nMid = 0;
while (nMid < nSize)
{
nMid += 5000000 / sizeof(T);
if (nMid > nSize)
nMid = nSize;
v.resize(nMid);
for (; i < nMid; i++)
Unserialize(is, v[i], nType, nVersion);
}
}
template<typename Stream, typename T, typename A>
inline void Unserialize(Stream& is, std::vector<T, A>& v, int nType, int nVersion)
{
Unserialize_impl(is, v, nType, nVersion, T());
}
/**
* others derived from vector
*/
inline unsigned int GetSerializeSize(const CScript& v, int nType, int nVersion)
{
return GetSerializeSize((const std::vector<unsigned char>&)v, nType, nVersion);
}
template<typename Stream>
void Serialize(Stream& os, const CScript& v, int nType, int nVersion)
{
Serialize(os, (const std::vector<unsigned char>&)v, nType, nVersion);
}
template<typename Stream>
void Unserialize(Stream& is, CScript& v, int nType, int nVersion)
{
Unserialize(is, (std::vector<unsigned char>&)v, nType, nVersion);
}
/**
* pair
*/
template<typename K, typename T>
unsigned int GetSerializeSize(const std::pair<K, T>& item, int nType, int nVersion)
{
return GetSerializeSize(item.first, nType, nVersion) + GetSerializeSize(item.second, nType, nVersion);
}
template<typename Stream, typename K, typename T>
void Serialize(Stream& os, const std::pair<K, T>& item, int nType, int nVersion)
{
Serialize(os, item.first, nType, nVersion);
Serialize(os, item.second, nType, nVersion);
}
template<typename Stream, typename K, typename T>
void Unserialize(Stream& is, std::pair<K, T>& item, int nType, int nVersion)
{
Unserialize(is, item.first, nType, nVersion);
Unserialize(is, item.second, nType, nVersion);
}
/**
* map
*/
template<typename K, typename T, typename Pred, typename A>
unsigned int GetSerializeSize(const std::map<K, T, Pred, A>& m, int nType, int nVersion)
{
unsigned int nSize = GetSizeOfCompactSize(m.size());
for (typename std::map<K, T, Pred, A>::const_iterator mi = m.begin(); mi != m.end(); ++mi)
nSize += GetSerializeSize((*mi), nType, nVersion);
return nSize;
}
template<typename Stream, typename K, typename T, typename Pred, typename A>
void Serialize(Stream& os, const std::map<K, T, Pred, A>& m, int nType, int nVersion)
{
WriteCompactSize(os, m.size());
for (typename std::map<K, T, Pred, A>::const_iterator mi = m.begin(); mi != m.end(); ++mi)
Serialize(os, (*mi), nType, nVersion);
}
template<typename Stream, typename K, typename T, typename Pred, typename A>
void Unserialize(Stream& is, std::map<K, T, Pred, A>& m, int nType, int nVersion)
{
m.clear();
unsigned int nSize = ReadCompactSize(is);
typename std::map<K, T, Pred, A>::iterator mi = m.begin();
for (unsigned int i = 0; i < nSize; i++)
{
std::pair<K, T> item;
Unserialize(is, item, nType, nVersion);
mi = m.insert(mi, item);
}
}
/**
* set
*/
template<typename K, typename Pred, typename A>
unsigned int GetSerializeSize(const std::set<K, Pred, A>& m, int nType, int nVersion)
{
unsigned int nSize = GetSizeOfCompactSize(m.size());
for (typename std::set<K, Pred, A>::const_iterator it = m.begin(); it != m.end(); ++it)
nSize += GetSerializeSize((*it), nType, nVersion);
return nSize;
}
template<typename Stream, typename K, typename Pred, typename A>
void Serialize(Stream& os, const std::set<K, Pred, A>& m, int nType, int nVersion)
{
WriteCompactSize(os, m.size());
for (typename std::set<K, Pred, A>::const_iterator it = m.begin(); it != m.end(); ++it)
Serialize(os, (*it), nType, nVersion);
}
template<typename Stream, typename K, typename Pred, typename A>
void Unserialize(Stream& is, std::set<K, Pred, A>& m, int nType, int nVersion)
{
m.clear();
unsigned int nSize = ReadCompactSize(is);
typename std::set<K, Pred, A>::iterator it = m.begin();
for (unsigned int i = 0; i < nSize; i++)
{
K key;
Unserialize(is, key, nType, nVersion);
it = m.insert(it, key);
}
}
/**
* Support for ADD_SERIALIZE_METHODS and READWRITE macro
*/
struct CSerActionSerialize
{
bool ForRead() const { return false; }
};
struct CSerActionUnserialize
{
bool ForRead() const { return true; }
};
template<typename Stream, typename T>
inline void SerReadWrite(Stream& s, const T& obj, int nType, int nVersion, CSerActionSerialize ser_action)
{
::Serialize(s, obj, nType, nVersion);
}
template<typename Stream, typename T>
inline void SerReadWrite(Stream& s, T& obj, int nType, int nVersion, CSerActionUnserialize ser_action)
{
::Unserialize(s, obj, nType, nVersion);
}
class CSizeComputer
{
protected:
size_t nSize;
public:
int nType;
int nVersion;
CSizeComputer(int nTypeIn, int nVersionIn) : nSize(0), nType(nTypeIn), nVersion(nVersionIn) {}
CSizeComputer& write(const char *psz, size_t nSize)
{
this->nSize += nSize;
return *this;
}
template<typename T>
CSizeComputer& operator<<(const T& obj)
{
::Serialize(*this, obj, nType, nVersion);
return (*this);
}
size_t size() const {
return nSize;
}
};
#endif // BITCOIN_SERIALIZE_H
+3 -8
View File
@@ -355,7 +355,6 @@ extern "C" int scanhash_skeincoin(int thr_id, uint32_t *pdata, const uint32_t *p
const int swap = 1;
bool sm5 = (device_sm[device_map[thr_id]] >= 500);
bool checkSecnonce = (have_stratum || have_longpoll) && !sm5;
uint32_t throughput = device_intensity(thr_id, __func__, 1U << 20);
throughput = min(throughput, (max_nonce - first_nonce));
@@ -417,7 +416,7 @@ extern "C" int scanhash_skeincoin(int thr_id, uint32_t *pdata, const uint32_t *p
if (vhash64[7] <= ptarget[7] && fulltest(vhash64, ptarget)) {
int res = 1;
uint8_t num = res;
if (checkSecnonce) {
if (!sm5) {
secNonce = cuda_check_hash_suppl(thr_id, throughput, pdata[19], d_hash[thr_id], num);
}
while (secNonce != 0 && res < 2) /* todo: up to 6 */
@@ -430,10 +429,7 @@ extern "C" int scanhash_skeincoin(int thr_id, uint32_t *pdata, const uint32_t *p
res++;
}
num++;
//if (checkSecnonce)
// secNonce = cuda_check_hash_suppl(thr_id, throughput, pdata[19], d_hash[thr_id], num);
//else
break; // only one secNonce...
secNonce = cuda_check_hash_suppl(thr_id, throughput, pdata[19], d_hash[thr_id], num);
}
if (res > 1 && opt_debug)
applog(LOG_BLUE, "GPU #%d: %d/%d valid nonces !!!", device_map[thr_id], res, (int)num);
@@ -445,8 +441,7 @@ extern "C" int scanhash_skeincoin(int thr_id, uint32_t *pdata, const uint32_t *p
}
}
if ((uint64_t) throughput + pdata[19] > max_nonce) {
//applog(LOG_DEBUG, "done... max=%u", max_nonce);
if ((uint64_t) pdata[19] + throughput > max_nonce) {
*hashes_done = pdata[19] - first_nonce;
pdata[19] = max_nonce;
break;
+1 -1
View File
@@ -101,7 +101,7 @@ int scanhash_skein2(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
}
}
if ((uint64_t) throughput + pdata[19] > max_nonce) {
if ((uint64_t) pdata[19] + throughput > max_nonce) {
*hashes_done = pdata[19] - first_nonce;
pdata[19] = max_nonce;
break;
+7 -10
View File
@@ -1770,9 +1770,9 @@ void do_gpu_tests(void)
//memcpy(buf, zrtest, 80);
//scanhash_zr5(0, (uint32_t*)buf, tgt, zrtest[19]+1, &done);
//struct timeval tv;
//memset(buf, 0, sizeof buf);
//scanhash_scrypt_jane(0, (uint32_t*)buf, tgt, NULL, 1, &done, &tv, &tv);
struct timeval tv;
memset(buf, 0, sizeof buf);
scanhash_scrypt_jane(0, (uint32_t*)buf, tgt, NULL, 1, &done, &tv, &tv);
memset(buf, 0, sizeof buf);
scanhash_x11(0, (uint32_t*)buf, tgt, 1, &done);
@@ -1814,9 +1814,6 @@ void print_hash_tests(void)
blake256hash(&hash[0], &buf[0], 14);
printpfx("blake", hash);
c11hash(&hash[0], &buf[0]);
printpfx("c11", hash);
deephash(&hash[0], &buf[0]);
printpfx("deep", hash);
@@ -1838,7 +1835,7 @@ void print_hash_tests(void)
keccak256_hash(&hash[0], &buf[0]);
printpfx("keccak", hash);
luffa_hash(&hash[0], &buf[0]);
doomhash(&hash[0], &buf[0]);
printpfx("luffa", hash);
lyra2_hash(&hash[0], &buf[0]);
@@ -1856,6 +1853,9 @@ void print_hash_tests(void)
pentablakehash(&hash[0], &buf[0]);
printpfx("pentablake", hash);
pluckhash((uint32_t*)&hash[0], (uint32_t*)&buf[0], scratchbuf, 128);
printpfx("pluck", hash);
quarkhash(&hash[0], &buf[0]);
printpfx("quark", hash);
@@ -1877,9 +1877,6 @@ void print_hash_tests(void)
s3hash(&hash[0], &buf[0]);
printpfx("S3", hash);
whirlxHash(&hash[0], &buf[0]);
printpfx("whirlpoolx", hash);
x11hash(&hash[0], &buf[0]);
printpfx("X11", hash);
-242
View File
@@ -1,242 +0,0 @@
extern "C"
{
#include "sph/sph_blake.h"
#include "sph/sph_bmw.h"
#include "sph/sph_groestl.h"
#include "sph/sph_skein.h"
#include "sph/sph_jh.h"
#include "sph/sph_keccak.h"
#include "sph/sph_luffa.h"
#include "sph/sph_cubehash.h"
#include "sph/sph_shavite.h"
#include "sph/sph_simd.h"
#include "sph/sph_echo.h"
}
#include "miner.h"
#include "cuda_helper.h"
#include <stdio.h>
#include <memory.h>
static uint32_t *d_hash[MAX_GPUS];
extern void quark_blake512_cpu_init(int thr_id, uint32_t threads);
extern void quark_blake512_cpu_setBlock_80(int thr_id, uint32_t *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash);
extern void quark_bmw512_cpu_init(int thr_id, uint32_t threads);
extern void quark_bmw512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_groestl512_cpu_init(int thr_id, uint32_t threads);
extern void quark_groestl512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_doublegroestl512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_skein512_cpu_init(int thr_id, uint32_t threads);
extern void quark_skein512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_keccak512_cpu_init(int thr_id, uint32_t threads);
extern void quark_keccak512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern int x11_simd512_cpu_init(int thr_id, uint32_t threads);
extern void x11_simd512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_echo512_cpu_init(int thr_id, uint32_t threads);
extern void x11_echo512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_compactTest_cpu_init(int thr_id, uint32_t threads);
extern void quark_compactTest_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *inpHashes,
uint32_t *d_noncesTrue, size_t *nrmTrue, uint32_t *d_noncesFalse, size_t *nrmFalse, int order);
// Flax/C11 CPU Hash
extern "C" void c11hash(void *output, const void *input)
{
// blake1-bmw2-grs3-skein4-jh5-keccak6-luffa7-cubehash8-shavite9-simd10-echo11
sph_blake512_context ctx_blake;
sph_bmw512_context ctx_bmw;
sph_groestl512_context ctx_groestl;
sph_jh512_context ctx_jh;
sph_keccak512_context ctx_keccak;
sph_skein512_context ctx_skein;
sph_luffa512_context ctx_luffa;
sph_cubehash512_context ctx_cubehash;
sph_shavite512_context ctx_shavite;
sph_simd512_context ctx_simd;
sph_echo512_context ctx_echo;
unsigned char hash[128];
memset(hash, 0, sizeof hash);
sph_blake512_init(&ctx_blake);
sph_blake512 (&ctx_blake, input, 80);
sph_blake512_close(&ctx_blake, (void*) hash);
sph_bmw512_init(&ctx_bmw);
sph_bmw512 (&ctx_bmw, (const void*) hash, 64);
sph_bmw512_close(&ctx_bmw, (void*) hash);
sph_groestl512_init(&ctx_groestl);
sph_groestl512 (&ctx_groestl, (const void*) hash, 64);
sph_groestl512_close(&ctx_groestl, (void*) hash);
sph_jh512_init(&ctx_jh);
sph_jh512 (&ctx_jh, (const void*) hash, 64);
sph_jh512_close(&ctx_jh, (void*) hash);
sph_keccak512_init(&ctx_keccak);
sph_keccak512 (&ctx_keccak, (const void*) hash, 64);
sph_keccak512_close(&ctx_keccak, (void*) hash);
sph_skein512_init(&ctx_skein);
sph_skein512 (&ctx_skein, (const void*) hash, 64);
sph_skein512_close(&ctx_skein, (void*) hash);
sph_luffa512_init(&ctx_luffa);
sph_luffa512 (&ctx_luffa, (const void*) hash, 64);
sph_luffa512_close (&ctx_luffa, (void*) hash);
sph_cubehash512_init(&ctx_cubehash);
sph_cubehash512 (&ctx_cubehash, (const void*) hash, 64);
sph_cubehash512_close(&ctx_cubehash, (void*) hash);
sph_shavite512_init(&ctx_shavite);
sph_shavite512 (&ctx_shavite, (const void*) hash, 64);
sph_shavite512_close(&ctx_shavite, (void*) hash);
sph_simd512_init(&ctx_simd);
sph_simd512 (&ctx_simd, (const void*) hash, 64);
sph_simd512_close(&ctx_simd, (void*) hash);
sph_echo512_init(&ctx_echo);
sph_echo512 (&ctx_echo, (const void*) hash, 64);
sph_echo512_close(&ctx_echo, (void*) hash);
memcpy(output, hash, 32);
}
#ifdef _DEBUG
#define TRACE(algo) { \
if (max_nonce == 1 && pdata[19] <= 1) { \
uint32_t* debugbuf = NULL; \
cudaMallocHost(&debugbuf, 8*sizeof(uint32_t)); \
cudaMemcpy(debugbuf, d_hash[thr_id], 8*sizeof(uint32_t), cudaMemcpyDeviceToHost); \
printf("X11 %s %08x %08x %08x %08x...\n", algo, swab32(debugbuf[0]), swab32(debugbuf[1]), \
swab32(debugbuf[2]), swab32(debugbuf[3])); \
cudaFreeHost(debugbuf); \
} \
}
#else
#define TRACE(algo) {}
#endif
static bool init[MAX_GPUS] = { 0 };
extern "C" int scanhash_c11(int thr_id, uint32_t *pdata, const uint32_t *ptarget, uint32_t max_nonce, unsigned long *hashes_done)
{
const uint32_t first_nonce = pdata[19];
int intensity = (device_sm[device_map[thr_id]] >= 500 && !is_windows()) ? 20 : 19;
uint32_t throughput = device_intensity(thr_id, __func__, 1U << intensity); // 19=256*256*8;
throughput = min(throughput, max_nonce - first_nonce);
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x5;
if (!init[thr_id])
{
cudaSetDevice(device_map[thr_id]);
quark_blake512_cpu_init(thr_id, throughput);
quark_bmw512_cpu_init(thr_id, throughput);
quark_groestl512_cpu_init(thr_id, throughput);
quark_skein512_cpu_init(thr_id, throughput);
quark_keccak512_cpu_init(thr_id, throughput);
quark_jh512_cpu_init(thr_id, throughput);
x11_luffaCubehash512_cpu_init(thr_id, throughput);
x11_shavite512_cpu_init(thr_id, throughput);
x11_echo512_cpu_init(thr_id, throughput);
if (x11_simd512_cpu_init(thr_id, throughput) != 0) {
return 0;
}
CUDA_CALL_OR_RET_X(cudaMalloc(&d_hash[thr_id], 64 * throughput), 0); // why 64 ?
cuda_check_cpu_init(thr_id, throughput);
init[thr_id] = true;
}
uint32_t endiandata[20];
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
int order = 0;
uint32_t foundNonce;
// Hash with CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
TRACE("blake :");
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("bmw :");
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("groestl:");
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("jh512 :");
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("keccak :");
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("skein :");
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
TRACE("luffa+c:");
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("shavite:");
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("simd :");
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("echo => ");
foundNonce = cuda_check_hash(thr_id, throughput, pdata[19], d_hash[thr_id]);
if (foundNonce != UINT32_MAX)
{
const uint32_t Htarg = ptarget[7];
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
c11hash(vhash64, endiandata);
if (vhash64[7] <= Htarg && fulltest(vhash64, ptarget)) {
int res = 1;
// check if there was some other ones...
uint32_t secNonce = cuda_check_hash_suppl(thr_id, throughput, pdata[19], d_hash[thr_id], 1);
*hashes_done = pdata[19] - first_nonce + throughput;
if (secNonce != 0) {
pdata[21] = secNonce;
res++;
}
pdata[19] = foundNonce;
return res;
} else {
applog(LOG_WARNING, "GPU #%d: result for %08x does not validate on CPU!", device_map[thr_id], foundNonce);
pdata[19] = foundNonce + 1;
}
}
pdata[19] += throughput;
} while (pdata[19] < max_nonce && !work_restart[thr_id].restart);
*hashes_done = pdata[19] - first_nonce + 1;
return 0;
}
File diff suppressed because it is too large. Load diff
+2 -3
View File
@@ -43,7 +43,7 @@ extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -201,7 +201,7 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata,
TRACE("jh512 :");
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("keccak :");
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("luffa+c:");
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
TRACE("shavite:");
@@ -231,7 +231,6 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata,
return res;
} else {
applog(LOG_WARNING, "GPU #%d: result for %08x does not validate on CPU!", device_map[thr_id], foundNonce);
pdata[19] = foundNonce + 1;
}
}
+2 -2
View File
@@ -45,7 +45,7 @@ extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -200,7 +200,7 @@ extern "C" int scanhash_x13(int thr_id, uint32_t *pdata,
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
-585
View File
@@ -1,585 +0,0 @@
/*
* Built on cbuchner1's implementation, actual hashing code
* based on sphlib 3.0
*/
#include <stdio.h>
#include <memory.h>
#define threadsPerBlock 1024
#include "cuda_helper.h"
__constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
__constant__ uint64_t c_xtra[8];
__constant__ uint64_t c_tmp[72];
__constant__ uint64_t pTarget[4];
static uint32_t *h_wxnounce[MAX_GPUS] = { 0 };
static uint32_t *d_WXNonce[MAX_GPUS] = { 0 };
/**
* Whirlpool CUDA kernel implementation.
*
* ==========================(LICENSE BEGIN)============================
*
* Copyright (c) 2014 djm34 & tpruvot & SP & Provos Alexis
*
* Permission is hereby granted, free of charge, to any person obtaining
* a copy of this software and associated documentation files (the
* "Software"), to deal in the Software without restriction, including
* without limitation the rights to use, copy, modify, merge, publish,
* distribute, sublicense, and/or sell copies of the Software, and to
* permit persons to whom the Software is furnished to do so, subject to
* the following conditions:
*
* The above copyright notice and this permission notice shall be
* included in all copies or substantial portions of the Software.
*
* THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
* EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
* MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
* IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
* CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
* TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
* SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
*
* ===========================(LICENSE END)=============================
* @author djm34
* @author tpruvot
* @author SP
* @author Provos Alexis
*/
__constant__ __align__(64) uint64_t mixTob0Tox[256];
const uint64_t plain_T0[256]= {
0xD83078C018601818,0x2646AF05238C2323,0xB891F97EC63FC6C6,0xFBCD6F13E887E8E8,0xCB13A14C87268787,0x116D62A9B8DAB8B8,0x0902050801040101,0x0D9E6E424F214F4F,0x9B6CEEAD36D83636,
0xFF510459A6A2A6A6,0x0CB9BDDED26FD2D2,0x0EF706FBF5F3F5F5,0x96F280EF79F97979,0x30DECE5F6FA16F6F,0x6D3FEFFC917E9191,0xF8A407AA52555252,0x47C0FD27609D6060,0x35657689BCCABCBC,
0x372BCDAC9B569B9B,0x8A018C048E028E8E,0xD25B1571A3B6A3A3,0x6C183C600C300C0C,0x84F68AFF7BF17B7B,0x806AE1B535D43535,0xF53A69E81D741D1D,0xB3DD4753E0A7E0E0,0x21B3ACF6D77BD7D7,
0x9C99ED5EC22FC2C2,0x435C966D2EB82E2E,0x29967A624B314B4B,0x5DE121A3FEDFFEFE,0xD5AE168257415757,0xBD2A41A815541515,0xE8EEB69F77C17777,0x926EEBA537DC3737,0x9ED7567BE5B3E5E5,
0x1323D98C9F469F9F,0x23FD17D3F0E7F0F0,0x20947F6A4A354A4A,0x44A9959EDA4FDADA,0xA2B025FA587D5858,0xCF8FCA06C903C9C9,0x7C528D5529A42929,0x5A1422500A280A0A,0x507F4FE1B1FEB1B1,
0xC95D1A69A0BAA0A0,0x14D6DA7F6BB16B6B,0xD917AB5C852E8585,0x3C677381BDCEBDBD,0x8FBA34D25D695D5D,0x9020508010401010,0x07F503F3F4F7F4F4,0xDD8BC016CB0BCBCB,0xD37CC6ED3EF83E3E,
0x2D0A112805140505,0x78CEE61F67816767,0x97D55373E4B7E4E4,0x024EBB25279C2727,0x7382583241194141,0xA70B9D2C8B168B8B,0xF6530151A7A6A7A7,0xB2FA94CF7DE97D7D,0x4937FBDC956E9595,
0x56AD9F8ED847D8D8,0x70EB308BFBCBFBFB,0xCDC17123EE9FEEEE,0xBBF891C77CED7C7C,0x71CCE31766856666,0x7BA78EA6DD53DDDD,0xAF2E4BB8175C1717,0x458E460247014747,0x1A21DC849E429E9E,
0xD489C51ECA0FCACA,0x585A99752DB42D2D,0x2E637991BFC6BFBF,0x3F0E1B38071C0707,0xAC472301AD8EADAD,0xB0B42FEA5A755A5A,0xEF1BB56C83368383,0xB666FF8533CC3333,0x5CC6F23F63916363,
0x12040A1002080202,0x93493839AA92AAAA,0xDEE2A8AF71D97171,0xC68DCF0EC807C8C8,0xD1327DC819641919,0x3B92707249394949,0x5FAF9A86D943D9D9,0x31F91DC3F2EFF2F2,0xA8DB484BE3ABE3E3,
0xB9B62AE25B715B5B,0xBC0D9234881A8888,0x3E29C8A49A529A9A,0x0B4CBE2D26982626,0xBF64FA8D32C83232,0x597D4AE9B0FAB0B0,0xF2CF6A1BE983E9E9,0x771E33780F3C0F0F,0x33B7A6E6D573D5D5,
0xF41DBA74803A8080,0x27617C99BEC2BEBE,0xEB87DE26CD13CDCD,0x8968E4BD34D03434,0x3290757A483D4848,0x54E324ABFFDBFFFF,0x8DF48FF77AF57A7A,0x643DEAF4907A9090,0x9DBE3EC25F615F5F,
0x3D40A01D20802020,0x0FD0D56768BD6868,0xCA3472D01A681A1A,0xB7412C19AE82AEAE,0x7D755EC9B4EAB4B4,0xCEA8199A544D5454,0x7F3BE5EC93769393,0x2F44AA0D22882222,0x63C8E907648D6464,
0x2AFF12DBF1E3F1F1,0xCCE6A2BF73D17373,0x82245A9012481212,0x7A805D3A401D4040,0x4810284008200808,0x959BE856C32BC3C3,0xDFC57B33EC97ECEC,0x4DAB9096DB4BDBDB,0xC05F1F61A1BEA1A1,
0x9107831C8D0E8D8D,0xC87AC9F53DF43D3D,0x5B33F1CC97669797,0x0000000000000000,0xF983D436CF1BCFCF,0x6E5687452BAC2B2B,0xE1ECB39776C57676,0xE619B06482328282,0x28B1A9FED67FD6D6,
0xC33677D81B6C1B1B,0x74775BC1B5EEB5B5,0xBE432911AF86AFAF,0x1DD4DF776AB56A6A,0xEAA00DBA505D5050,0x578A4C1245094545,0x38FB18CBF3EBF3F3,0xAD60F09D30C03030,0xC4C3742BEF9BEFEF,
0xDA7EC3E53FFC3F3F,0xC7AA1C9255495555,0xDB591079A2B2A2A2,0xE9C96503EA8FEAEA,0x6ACAEC0F65896565,0x036968B9BAD2BABA,0x4A5E93652FBC2F2F,0x8E9DE74EC027C0C0,0x60A181BEDE5FDEDE,
0xFC386CE01C701C1C,0x46E72EBBFDD3FDFD,0x1F9A64524D294D4D,0x7639E0E492729292,0xFAEABC8F75C97575,0x360C1E3006180606,0xAE0998248A128A8A,0x4B7940F9B2F2B2B2,0x85D15963E6BFE6E6,
0x7E1C36700E380E0E,0xE73E63F81F7C1F1F,0x55C4F73762956262,0x3AB5A3EED477D4D4,0x814D3229A89AA8A8,0x5231F4C496629696,0x62EF3A9BF9C3F9F9,0xA397F666C533C5C5,0x104AB13525942525,
0xABB220F259795959,0xD015AE54842A8484,0xC5E4A7B772D57272,0xEC72DDD539E43939,0x1698615A4C2D4C4C,0x94BC3BCA5E655E5E,0x9FF085E778FD7878,0xE570D8DD38E03838,0x980586148C0A8C8C,
0x17BFB2C6D163D1D1,0xE4570B41A5AEA5A5,0xA1D94D43E2AFE2E2,0x4EC2F82F61996161,0x427B45F1B3F6B3B3,0x3442A51521842121,0x0825D6949C4A9C9C,0xEE3C66F01E781E1E,0x6186522243114343,
0xB193FC76C73BC7C7,0x4FE52BB3FCD7FCFC,0x2408142004100404,0xE3A208B251595151,0x252FC7BC995E9999,0x22DAC44F6DA96D6D,0x651A39680D340D0D,0x79E93583FACFFAFA,0x69A384B6DF5BDFDF,
0xA9FC9BD77EE57E7E,0x1948B43D24902424,0xFE76D7C53BEC3B3B,0x9A4B3D31AB96ABAB,0xF081D13ECE1FCECE,0x9922558811441111,0x8303890C8F068F8F,0x049C6B4A4E254E4E,0x667351D1B7E6B7B7,
0xE0CB600BEB8BEBEB,0xC178CCFD3CF03C3C,0xFD1FBF7C813E8181,0x4035FED4946A9494,0x1CF30CEBF7FBF7F7,0x186F67A1B9DEB9B9,0x8B265F98134C1313,0x51589C7D2CB02C2C,0x05BBB8D6D36BD3D3,
0x8CD35C6BE7BBE7E7,0x39DCCB576EA56E6E,0xAA95F36EC437C4C4,0x1B060F18030C0303,0xDCAC138A56455656,0x5E88491A440D4444,0xA0FE9EDF7FE17F7F,0x884F3721A99EA9A9,0x6754824D2AA82A2A,
0x0A6B6DB1BBD6BBBB,0x879FE246C123C1C1,0xF1A602A253515353,0x72A58BAEDC57DCDC,0x531627580B2C0B0B,0x0127D39C9D4E9D9D,0x2BD8C1476CAD6C6C,0xA462F59531C43131,0xF3E8B98774CD7474,
0x15F109E3F6FFF6F6,0x4C8C430A46054646,0xA5452609AC8AACAC,0xB50F973C891E8989,0xB42844A014501414,0xBADF425BE1A3E1E1,0xA62C4EB016581616,0xF774D2CD3AE83A3A,0x06D2D06F69B96969,
0x41122D4809240909,0xD7E0ADA770DD7070,0x6F7154D9B6E2B6B6,0x1EBDB7CED067D0D0,0xD6C77E3BED93EDED,0xE285DB2ECC17CCCC,0x6884572A42154242,0x2C2DC2B4985A9898,0xED550E49A4AAA4A4,
0x7550885D28A02828,0x86B831DA5C6D5C5C,0x6BED3F93F8C7F8F8,0xC211A44486228686
};
/**
* Round constants.
*/
__constant__ uint64_t InitVector_RC[10];
const uint64_t plain_RC[10] = {
0x4F01B887E8C62318,0x52916F79F5D2A636,0x357B0CA38E9BBC60,0x57FE4B2EC2D7E01D,0xDA4AF09FE5377715,
0x856BA0B10A29C958,0x67053ECBF4105DBD,0xD8957DA78B4127E4,0x9E4717DD667CEEFB,0x33835AAD07BF2DCA
};
/* ====================================================================== */
__device__ __forceinline__
static uint64_t ROUND_ELT(const uint64_t* sharedMemory, const uint64_t* __restrict__ in, const int i0, const int i1, const int i2, const int i3, const int i4, const int i5, const int i6, const int i7)
{
uint32_t* in32 = (uint32_t*)in;
return xor8( sharedMemory[__byte_perm(in32[(i0 << 1)], 0, 0x4440)],
sharedMemory[__byte_perm(in32[(i1 << 1)], 0, 0x4441) + 256],
sharedMemory[__byte_perm(in32[(i2 << 1)], 0, 0x4442) + 512],
sharedMemory[__byte_perm(in32[(i3 << 1)], 0, 0x4443) + 768],
sharedMemory[__byte_perm(in32[(i4 << 1) + 1], 0, 0x4440) + 1024],
sharedMemory[__byte_perm(in32[(i5 << 1) + 1], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(in32[(i6 << 1) + 1], 0, 0x4442) + 1536],
sharedMemory[__byte_perm(in32[(i7 << 1) + 1], 0, 0x4443) + 1792]);
}
#define TRANSFER(dst, src) { \
dst[0] = src ## 0; \
dst[1] = src ## 1; \
dst[2] = src ## 2; \
dst[3] = src ## 3; \
dst[4] = src ## 4; \
dst[5] = src ## 5; \
dst[6] = src ## 6; \
dst[7] = src ## 7; \
}
#define ROUND(table, in, out, c0, c1, c2, c3, c4, c5, c6, c7) { \
out ## 0 = xor1(ROUND_ELT(table, in, 0, 7, 6, 5, 4, 3, 2, 1), c0); \
out ## 1 = xor1(ROUND_ELT(table, in, 1, 0, 7, 6, 5, 4, 3, 2), c1); \
out ## 2 = xor1(ROUND_ELT(table, in, 2, 1, 0, 7, 6, 5, 4, 3), c2); \
out ## 3 = xor1(ROUND_ELT(table, in, 3, 2, 1, 0, 7, 6, 5, 4), c3); \
out ## 4 = xor1(ROUND_ELT(table, in, 4, 3, 2, 1, 0, 7, 6, 5), c4); \
out ## 5 = xor1(ROUND_ELT(table, in, 5, 4, 3, 2, 1, 0, 7, 6), c5); \
out ## 6 = xor1(ROUND_ELT(table, in, 6, 5, 4, 3, 2, 1, 0, 7), c6); \
out ## 7 = xor1(ROUND_ELT(table, in, 7, 6, 5, 4, 3, 2, 1, 0), c7); \
}
#define ROUND1(table, in, out, c) { \
out ## 0 = xor1(ROUND_ELT(table, in, 0, 7, 6, 5, 4, 3, 2, 1), c); \
out ## 1 = ROUND_ELT(table, in, 1, 0, 7, 6, 5, 4, 3, 2); \
out ## 2 = ROUND_ELT(table, in, 2, 1, 0, 7, 6, 5, 4, 3); \
out ## 3 = ROUND_ELT(table, in, 3, 2, 1, 0, 7, 6, 5, 4); \
out ## 4 = ROUND_ELT(table, in, 4, 3, 2, 1, 0, 7, 6, 5); \
out ## 5 = ROUND_ELT(table, in, 5, 4, 3, 2, 1, 0, 7, 6); \
out ## 6 = ROUND_ELT(table, in, 6, 5, 4, 3, 2, 1, 0, 7); \
out ## 7 = ROUND_ELT(table, in, 7, 6, 5, 4, 3, 2, 1, 0); \
}
#define ROUND_KSCHED(table, in, out, c) \
ROUND1(table, in, out, c) \
TRANSFER(in, out)
#define ROUND_WENC(table, in, key, out) \
ROUND(table, in, out, key[0], key[1], key[2],key[3], key[4], key[5], key[6], key[7]) \
TRANSFER(in, out)
static uint64_t* d_xtra[MAX_GPUS] = { 0 };
static uint64_t* d_tmp[MAX_GPUS] = { 0 };
__device__ __forceinline__
static void whirlpoolx_getShared(uint64_t* sharedMemory)
{
if (threadIdx.x < 256) {
sharedMemory[threadIdx.x] = mixTob0Tox[threadIdx.x];
sharedMemory[threadIdx.x+256] = ROTL64(sharedMemory[threadIdx.x], 8);
sharedMemory[threadIdx.x+512] = ROTL64(sharedMemory[threadIdx.x],16);
sharedMemory[threadIdx.x+768] = ROTL64(sharedMemory[threadIdx.x],24);
sharedMemory[threadIdx.x+1024] = ROTL64(sharedMemory[threadIdx.x],32);
sharedMemory[threadIdx.x+1280] = ROTR64(sharedMemory[threadIdx.x],24);
sharedMemory[threadIdx.x+1536] = ROTR64(sharedMemory[threadIdx.x],16);
sharedMemory[threadIdx.x+1792] = ROTR64(sharedMemory[threadIdx.x], 8);
}
__syncthreads();
}
__global__
void whirlpoolx_gpu_precompute(uint32_t threads, uint64_t* d_xtra, uint64_t* d_tmp)
{
__shared__ uint64_t sharedMemory[2048];
whirlpoolx_getShared(sharedMemory);
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint64_t n[8];
uint64_t h[8] = { 0 };
#pragma unroll 8
for (int i=0; i<8; i++) {
n[i] = c_PaddedMessage80[i]; // read data
}
//#pragma unroll 10
for (unsigned int r=0; r < 10; r++) {
uint64_t tmp0, tmp1, tmp2, tmp3, tmp4, tmp5, tmp6, tmp7;
ROUND_KSCHED(sharedMemory, h, tmp, InitVector_RC[r]);
ROUND_WENC(sharedMemory, n, h, tmp);
}
#pragma unroll 8
for (int i=0; i < 8; i++) {
h[i] = xor1(n[i],c_PaddedMessage80[i]);
}
if(threadIdx.x==0)d_xtra[threadIdx.x]=h[1];
uint64_t atLastCalc=xor1(h[3],h[5]);
//////////////////////////////////
n[0] = c_PaddedMessage80[8]; //read data
n[1] = c_PaddedMessage80[9]; //whirlpool
n[2] = 0x0000000000000080; //whirlpool
n[3] = 0;
n[4] = 0;
n[5] = 0;
n[6] = 0;
n[7] = 0x8002000000000000;
n[0] = xor1(n[0],h[0]);
n[2] = xor1(n[2],h[2]); n[3] = h[3];
n[4] = h[4]; n[5] = h[5];
n[6] = h[6]; n[7] = xor1(n[7],h[7]);
uint64_t tmp[8];
tmp[0] = xor1(ROUND_ELT(sharedMemory, h, 0, 7, 6, 5, 4, 3, 2, 1),InitVector_RC[0]);
tmp[1] = ROUND_ELT(sharedMemory, h, 1, 0, 7, 6, 5, 4, 3, 2);
tmp[2] = ROUND_ELT(sharedMemory, h, 2, 1, 0, 7, 6, 5, 4, 3);
tmp[3] = ROUND_ELT(sharedMemory, h, 3, 2, 1, 0, 7, 6, 5, 4);
tmp[4] = ROUND_ELT(sharedMemory, h, 4, 3, 2, 1, 0, 7, 6, 5);
tmp[5] = ROUND_ELT(sharedMemory, h, 5, 4, 3, 2, 1, 0, 7, 6);
tmp[6] = ROUND_ELT(sharedMemory, h, 6, 5, 4, 3, 2, 1, 0, 7);
tmp[7] = ROUND_ELT(sharedMemory, h, 7, 6, 5, 4, 3, 2, 1, 0);
uint64_t tmp2[8];
uint32_t* n32 = (uint32_t*)n;
tmp2[0]=xor8( sharedMemory[__byte_perm(n32[ 0], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[14], 0, 0x4441) + 256],
sharedMemory[__byte_perm(n32[12], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[10], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[ 9], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[ 7], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(n32[ 5], 0, 0x4442) + 1536] ,tmp[0]);
tmp2[1]=xor8( tmp[1] ,sharedMemory[__byte_perm(n32[ 0], 0, 0x4441) + 256],
sharedMemory[__byte_perm(n32[14], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[12], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[11], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[ 9], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(n32[ 7], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[ 5], 0, 0x4443) + 1792]);
tmp2[2]=xor8( sharedMemory[__byte_perm(n32[ 4], 0, 0x4440)] ,tmp[2] ,
sharedMemory[__byte_perm(n32[ 0], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[14], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[13], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[11], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(n32[ 9], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[ 7], 0, 0x4443) + 1792]);
tmp2[3]=xor8( sharedMemory[__byte_perm(n32[ 6], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[ 4], 0, 0x4441) + 256],
tmp[3] ,sharedMemory[__byte_perm(n32[ 0], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[15], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[13], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(n32[11], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[ 9], 0, 0x4443) + 1792]);
tmp2[4]=xor8( sharedMemory[__byte_perm(n32[ 8], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[ 6], 0, 0x4441) + 256] ,
sharedMemory[__byte_perm(n32[ 4], 0, 0x4442) + 512] ,tmp[4] ,
sharedMemory[__byte_perm(n32[ 1], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[15], 0, 0x4441) + 1280] ,
sharedMemory[__byte_perm(n32[13], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[11], 0, 0x4443) + 1792]);
tmp2[5]=xor8( sharedMemory[__byte_perm(n32[10], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[ 8], 0, 0x4441) + 256],
sharedMemory[__byte_perm(n32[ 6], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[ 4], 0, 0x4443) + 768],
tmp[5] ,sharedMemory[__byte_perm(n32[ 1], 0, 0x4441) + 1280],
sharedMemory[__byte_perm(n32[15], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[13], 0, 0x4443) + 1792]);
tmp2[6]=xor8( sharedMemory[__byte_perm(n32[12], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[10], 0, 0x4441) + 256],
sharedMemory[__byte_perm(n32[ 8], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[ 6], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[ 5], 0, 0x4440) + 1024] ,tmp[6],
sharedMemory[__byte_perm(n32[ 1], 0, 0x4442) + 1536] ,sharedMemory[__byte_perm(n32[15], 0, 0x4443) + 1792]);
tmp2[7]=xor8( sharedMemory[__byte_perm(n32[14], 0, 0x4440)] ,sharedMemory[__byte_perm(n32[12], 0, 0x4441) + 256],
sharedMemory[__byte_perm(n32[10], 0, 0x4442) + 512] ,sharedMemory[__byte_perm(n32[ 8], 0, 0x4443) + 768],
sharedMemory[__byte_perm(n32[ 7], 0, 0x4440) + 1024] ,sharedMemory[__byte_perm(n32[ 5], 0, 0x4441) + 1280],
tmp[7] ,sharedMemory[__byte_perm(n32[ 1], 0, 0x4443) + 1792]);
n[1] ^= h[1];
tmp2[1]^=sharedMemory[__byte_perm(n32[2], 0, 0x4440)];
tmp2[2]^=sharedMemory[__byte_perm(n32[2], 0, 0x4441) + 256];
tmp2[3]^=sharedMemory[__byte_perm(n32[2], 0, 0x4442) + 512];
tmp2[4]^=sharedMemory[__byte_perm(n32[2], 0, 0x4443) + 768];
d_tmp[threadIdx.x]=tmp2[threadIdx.x];
uint64_t tmp3[8];
tmp3[0] = xor1(ROUND_ELT(sharedMemory, tmp, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[1]);
tmp3[1] = ROUND_ELT(sharedMemory, tmp, 1, 0, 7, 6, 5, 4, 3, 2);
tmp3[2] = ROUND_ELT(sharedMemory, tmp, 2, 1, 0, 7, 6, 5, 4, 3);
tmp3[3] = ROUND_ELT(sharedMemory, tmp, 3, 2, 1, 0, 7, 6, 5, 4);
tmp3[4] = ROUND_ELT(sharedMemory, tmp, 4, 3, 2, 1, 0, 7, 6, 5);
tmp3[5] = ROUND_ELT(sharedMemory, tmp, 5, 4, 3, 2, 1, 0, 7, 6);
tmp3[6] = ROUND_ELT(sharedMemory, tmp, 6, 5, 4, 3, 2, 1, 0, 7);
tmp3[7] = ROUND_ELT(sharedMemory, tmp, 7, 6, 5, 4, 3, 2, 1, 0);
n32 = (uint32_t*)tmp2;
uint64_t tmp4[8];
tmp4[0]=( sharedMemory[__byte_perm(n32[ 9], 0, 0x4440) + 1024] ^sharedMemory[__byte_perm(n32[ 7], 0, 0x4441) + 1280]^
sharedMemory[__byte_perm(n32[ 5], 0, 0x4442) + 1536] ^sharedMemory[__byte_perm(n32[ 3], 0, 0x4443) + 1792]) ^tmp3[0];
tmp4[1]=(sharedMemory[__byte_perm(n32[ 2], 0, 0x4440)] ^sharedMemory[__byte_perm(n32[ 9], 0, 0x4441) + 1280]^
sharedMemory[__byte_perm(n32[ 7], 0, 0x4442) + 1536] ^sharedMemory[__byte_perm(n32[ 5], 0, 0x4443) + 1792]) ^tmp3[1];
tmp4[2]=(sharedMemory[__byte_perm(n32[ 4], 0, 0x4440)] ^sharedMemory[__byte_perm(n32[ 2], 0, 0x4441) + 256]^
sharedMemory[__byte_perm(n32[ 9], 0, 0x4442) + 1536] ^sharedMemory[__byte_perm(n32[ 7], 0, 0x4443) + 1792]) ^tmp3[2];
tmp4[3]=(sharedMemory[__byte_perm(n32[ 6], 0, 0x4440)] ^sharedMemory[__byte_perm(n32[ 4], 0, 0x4441) + 256]^
sharedMemory[__byte_perm(n32[ 2], 0, 0x4442) + 512] ^sharedMemory[__byte_perm(n32[ 9], 0, 0x4443) + 1792]) ^tmp3[3];
tmp4[4]=(sharedMemory[__byte_perm(n32[ 8], 0, 0x4440)] ^sharedMemory[__byte_perm(n32[ 6], 0, 0x4441) + 256]^
sharedMemory[__byte_perm(n32[ 4], 0, 0x4442) + 512] ^sharedMemory[__byte_perm(n32[ 2], 0, 0x4443) + 768]) ^tmp3[4];
tmp4[5]=(sharedMemory[__byte_perm(n32[ 8], 0, 0x4441) + 256] ^sharedMemory[__byte_perm(n32[ 6], 0, 0x4442) + 512]^
sharedMemory[__byte_perm(n32[ 4], 0, 0x4443) + 768] ^sharedMemory[__byte_perm(n32[ 3], 0, 0x4440) + 1024]) ^tmp3[5];
tmp4[6]=(sharedMemory[__byte_perm(n32[ 8], 0, 0x4442) + 512] ^sharedMemory[__byte_perm(n32[ 6], 0, 0x4443) + 768]^
sharedMemory[__byte_perm(n32[ 5], 0, 0x4440) + 1024] ^sharedMemory[__byte_perm(n32[ 3], 0, 0x4441) + 1280]) ^tmp3[6];
tmp4[7]=(sharedMemory[__byte_perm(n32[ 8], 0, 0x4443) + 768] ^sharedMemory[__byte_perm(n32[ 7], 0, 0x4440) + 1024]^
sharedMemory[__byte_perm(n32[ 5], 0, 0x4441) + 1280] ^sharedMemory[__byte_perm(n32[ 3], 0, 0x4442) + 1536]) ^tmp3[7];
d_tmp[threadIdx.x+16]=tmp4[threadIdx.x];
uint64_t tmp5[8];
tmp5[0] = xor1(ROUND_ELT(sharedMemory, tmp3, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[2]);
tmp5[1] = ROUND_ELT(sharedMemory, tmp3, 1, 0, 7, 6, 5, 4, 3, 2);
tmp5[2] = ROUND_ELT(sharedMemory, tmp3, 2, 1, 0, 7, 6, 5, 4, 3);
tmp5[3] = ROUND_ELT(sharedMemory, tmp3, 3, 2, 1, 0, 7, 6, 5, 4);
tmp5[4] = ROUND_ELT(sharedMemory, tmp3, 4, 3, 2, 1, 0, 7, 6, 5);
tmp5[5] = ROUND_ELT(sharedMemory, tmp3, 5, 4, 3, 2, 1, 0, 7, 6);
tmp5[6] = ROUND_ELT(sharedMemory, tmp3, 6, 5, 4, 3, 2, 1, 0, 7);
tmp5[7] = ROUND_ELT(sharedMemory, tmp3, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+8]=tmp5[threadIdx.x];
uint64_t tmp6[8];
tmp6[0] = xor1(ROUND_ELT(sharedMemory, tmp5, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[3]);
tmp6[1] = ROUND_ELT(sharedMemory, tmp5, 1, 0, 7, 6, 5, 4, 3, 2);
tmp6[2] = ROUND_ELT(sharedMemory, tmp5, 2, 1, 0, 7, 6, 5, 4, 3);
tmp6[3] = ROUND_ELT(sharedMemory, tmp5, 3, 2, 1, 0, 7, 6, 5, 4);
tmp6[4] = ROUND_ELT(sharedMemory, tmp5, 4, 3, 2, 1, 0, 7, 6, 5);
tmp6[5] = ROUND_ELT(sharedMemory, tmp5, 5, 4, 3, 2, 1, 0, 7, 6);
tmp6[6] = ROUND_ELT(sharedMemory, tmp5, 6, 5, 4, 3, 2, 1, 0, 7);
tmp6[7] = ROUND_ELT(sharedMemory, tmp5, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+24]=tmp6[threadIdx.x];
uint64_t tmp7[8];
tmp7[0] = xor1(ROUND_ELT(sharedMemory, tmp6, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[4]);
tmp7[1] = ROUND_ELT(sharedMemory, tmp6, 1, 0, 7, 6, 5, 4, 3, 2);
tmp7[2] = ROUND_ELT(sharedMemory, tmp6, 2, 1, 0, 7, 6, 5, 4, 3);
tmp7[3] = ROUND_ELT(sharedMemory, tmp6, 3, 2, 1, 0, 7, 6, 5, 4);
tmp7[4] = ROUND_ELT(sharedMemory, tmp6, 4, 3, 2, 1, 0, 7, 6, 5);
tmp7[5] = ROUND_ELT(sharedMemory, tmp6, 5, 4, 3, 2, 1, 0, 7, 6);
tmp7[6] = ROUND_ELT(sharedMemory, tmp6, 6, 5, 4, 3, 2, 1, 0, 7);
tmp7[7] = ROUND_ELT(sharedMemory, tmp6, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+32]=tmp7[threadIdx.x];
uint64_t tmp8[8];
tmp8[0] = xor1(ROUND_ELT(sharedMemory, tmp7, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[5]);
tmp8[1] = ROUND_ELT(sharedMemory, tmp7, 1, 0, 7, 6, 5, 4, 3, 2);
tmp8[2] = ROUND_ELT(sharedMemory, tmp7, 2, 1, 0, 7, 6, 5, 4, 3);
tmp8[3] = ROUND_ELT(sharedMemory, tmp7, 3, 2, 1, 0, 7, 6, 5, 4);
tmp8[4] = ROUND_ELT(sharedMemory, tmp7, 4, 3, 2, 1, 0, 7, 6, 5);
tmp8[5] = ROUND_ELT(sharedMemory, tmp7, 5, 4, 3, 2, 1, 0, 7, 6);
tmp8[6] = ROUND_ELT(sharedMemory, tmp7, 6, 5, 4, 3, 2, 1, 0, 7);
tmp8[7] = ROUND_ELT(sharedMemory, tmp7, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+40]=tmp8[threadIdx.x];
uint64_t tmp9[8];
tmp9[0] = xor1(ROUND_ELT(sharedMemory, tmp8, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[6]);
tmp9[1] = ROUND_ELT(sharedMemory, tmp8, 1, 0, 7, 6, 5, 4, 3, 2);
tmp9[2] = ROUND_ELT(sharedMemory, tmp8, 2, 1, 0, 7, 6, 5, 4, 3);
tmp9[3] = ROUND_ELT(sharedMemory, tmp8, 3, 2, 1, 0, 7, 6, 5, 4);
tmp9[4] = ROUND_ELT(sharedMemory, tmp8, 4, 3, 2, 1, 0, 7, 6, 5);
tmp9[5] = ROUND_ELT(sharedMemory, tmp8, 5, 4, 3, 2, 1, 0, 7, 6);
tmp9[6] = ROUND_ELT(sharedMemory, tmp8, 6, 5, 4, 3, 2, 1, 0, 7);
tmp9[7] = ROUND_ELT(sharedMemory, tmp8, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+48]=tmp9[threadIdx.x];
uint64_t tmp10[8];
tmp10[0] = xor1(ROUND_ELT(sharedMemory, tmp9, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[7]);
tmp10[1] = ROUND_ELT(sharedMemory, tmp9, 1, 0, 7, 6, 5, 4, 3, 2);
tmp10[2] = ROUND_ELT(sharedMemory, tmp9, 2, 1, 0, 7, 6, 5, 4, 3);
tmp10[3] = ROUND_ELT(sharedMemory, tmp9, 3, 2, 1, 0, 7, 6, 5, 4);
tmp10[4] = ROUND_ELT(sharedMemory, tmp9, 4, 3, 2, 1, 0, 7, 6, 5);
tmp10[5] = ROUND_ELT(sharedMemory, tmp9, 5, 4, 3, 2, 1, 0, 7, 6);
tmp10[6] = ROUND_ELT(sharedMemory, tmp9, 6, 5, 4, 3, 2, 1, 0, 7);
tmp10[7] = ROUND_ELT(sharedMemory, tmp9, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+56]=tmp10[threadIdx.x];
uint64_t tmp11[8];
tmp11[0] = xor1(ROUND_ELT(sharedMemory, tmp10, 0, 7, 6, 5, 4, 3, 2, 1), InitVector_RC[8]);
tmp11[1] = ROUND_ELT(sharedMemory, tmp10, 1, 0, 7, 6, 5, 4, 3, 2);
tmp11[2] = ROUND_ELT(sharedMemory, tmp10, 2, 1, 0, 7, 6, 5, 4, 3);
tmp11[3] = ROUND_ELT(sharedMemory, tmp10, 3, 2, 1, 0, 7, 6, 5, 4);
tmp11[4] = ROUND_ELT(sharedMemory, tmp10, 4, 3, 2, 1, 0, 7, 6, 5);
tmp11[5] = ROUND_ELT(sharedMemory, tmp10, 5, 4, 3, 2, 1, 0, 7, 6);
tmp11[6] = ROUND_ELT(sharedMemory, tmp10, 6, 5, 4, 3, 2, 1, 0, 7);
tmp11[7] = ROUND_ELT(sharedMemory, tmp10, 7, 6, 5, 4, 3, 2, 1, 0);
d_tmp[threadIdx.x+64]=tmp11[threadIdx.x];
if(threadIdx.x==1){
tmp[0]=ROUND_ELT(sharedMemory,tmp11, 3, 2, 1, 0, 7, 6, 5, 4);
tmp[1]=ROUND_ELT(sharedMemory,tmp11, 5, 4, 3, 2, 1, 0, 7, 6);
tmp[4] = xor3(tmp[0],tmp[1],atLastCalc);
d_xtra[threadIdx.x]=tmp[4];
}
}
}
__global__ __launch_bounds__(threadsPerBlock,2)
void whirlpoolx_gpu_hash(uint32_t threads, uint32_t startNounce, uint32_t *resNounce)
{
__shared__ uint64_t sharedMemory[2048];
whirlpoolx_getShared(sharedMemory);
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint64_t n[8];
uint64_t tmp[8];
uint32_t nounce = startNounce + thread;
n[1] = xor1(REPLACE_HIDWORD(c_PaddedMessage80[9], cuda_swab32(nounce)),c_xtra[0]);
uint32_t* n32 = (uint32_t*)&n[0];
n[0]=sharedMemory[__byte_perm(n32[3], 0, 0x4443) + 1792];
n[5]=sharedMemory[__byte_perm(n32[3], 0, 0x4440) + 1024];
n[6]=sharedMemory[__byte_perm(n32[3], 0, 0x4441) + 1280];
n[7]=sharedMemory[__byte_perm(n32[3], 0, 0x4442) + 1536];
n[0]=xor1(c_tmp[0],n[0]);
n[1]=c_tmp[1];
n[2]=c_tmp[2];
n[3]=c_tmp[3];
n[4]=c_tmp[4];
n[5]=xor1(c_tmp[5],n[5]);
n[6]=xor1(c_tmp[6],n[6]);
n[7]=xor1(c_tmp[7],n[7]);
tmp[0]=xor3(sharedMemory[__byte_perm(n32[10],0,0x4443)+768],sharedMemory[__byte_perm(n32[12],0,0x4442)+512],sharedMemory[__byte_perm(n32[14],0,0x4441)+256]);
tmp[1]=xor3(sharedMemory[__byte_perm(n32[11],0,0x4440)+1024],sharedMemory[__byte_perm(n32[12],0,0x4443)+768],sharedMemory[__byte_perm(n32[14],0,0x4442)+512]);
tmp[2]=xor3(sharedMemory[__byte_perm(n32[11],0,0x4441)+1280],sharedMemory[__byte_perm(n32[13],0,0x4440)+1024],sharedMemory[__byte_perm(n32[14],0,0x4443)+768]);
tmp[3]=xor3(sharedMemory[__byte_perm(n32[11],0,0x4442)+1536],sharedMemory[__byte_perm(n32[13],0,0x4441)+1280],sharedMemory[__byte_perm(n32[15],0,0x4440)+1024]);
tmp[4]=xor3(sharedMemory[__byte_perm(n32[11],0,0x4443)+1792],sharedMemory[__byte_perm(n32[13],0,0x4442)+1536],sharedMemory[__byte_perm(n32[15],0,0x4441)+1280]);
tmp[5]=xor3(sharedMemory[__byte_perm(n32[10],0,0x4440)],sharedMemory[__byte_perm(n32[13],0,0x4443)+1792],sharedMemory[__byte_perm(n32[15],0,0x4442)+1536]);
tmp[6]=xor3(sharedMemory[__byte_perm(n32[12],0,0x4440)],sharedMemory[__byte_perm(n32[10],0,0x4441)+256],sharedMemory[__byte_perm(n32[15],0,0x4443)+1792]);
tmp[7]=xor3(sharedMemory[__byte_perm(n32[14],0,0x4440)],sharedMemory[__byte_perm(n32[12],0,0x4441)+256],sharedMemory[__byte_perm(n32[10],0,0x4442)+ 512]);
tmp[0]=xor3(sharedMemory[__byte_perm(n32[ 0], 0, 0x4440)],tmp[0],c_tmp[0+16]);
tmp[1]=xor3(sharedMemory[__byte_perm(n32[ 0], 0, 0x4441) + 256],tmp[1],c_tmp[1+16]);
tmp[2]=xor3(sharedMemory[__byte_perm(n32[ 0], 0, 0x4442) + 512],tmp[2],c_tmp[2+16]);
tmp[3]=xor3(sharedMemory[__byte_perm(n32[ 0], 0, 0x4443) + 768],tmp[3],c_tmp[3+16]);
tmp[4]=xor3(sharedMemory[__byte_perm(n32[ 1], 0, 0x4440) + 1024],tmp[4],c_tmp[4+16]);
tmp[5]=xor3(sharedMemory[__byte_perm(n32[ 1], 0, 0x4441) + 1280],tmp[5],c_tmp[5+16]);
tmp[6]=xor3(sharedMemory[__byte_perm(n32[ 1], 0, 0x4442) + 1536],tmp[6],c_tmp[6+16]);
tmp[7]=xor3(sharedMemory[__byte_perm(n32[ 1], 0, 0x4443) + 1792],tmp[7],c_tmp[7+16]);
n[0]=tmp[0];
n[1]=tmp[1];
n[2]=tmp[2];
n[3]=tmp[3];
n[4]=tmp[4];
n[5]=tmp[5];
n[6]=tmp[6];
n[7]=tmp[7];
tmp[0] = xor1(ROUND_ELT(sharedMemory, n, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+8]);
tmp[1] = xor1(ROUND_ELT(sharedMemory, n, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+8]);
tmp[2] = xor1(ROUND_ELT(sharedMemory, n, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+8]);
tmp[3] = xor1(ROUND_ELT(sharedMemory, n, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+8]);
tmp[4] = xor1(ROUND_ELT(sharedMemory, n, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+8]);
tmp[5] = xor1(ROUND_ELT(sharedMemory, n, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+8]);
tmp[6] = xor1(ROUND_ELT(sharedMemory, n, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+8]);
tmp[7] = xor1(ROUND_ELT(sharedMemory, n, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+8]);
n[0] = xor1(ROUND_ELT(sharedMemory, tmp, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+24]);
n[1] = xor1(ROUND_ELT(sharedMemory, tmp, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+24]);
n[2] = xor1(ROUND_ELT(sharedMemory, tmp, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+24]);
n[3] = xor1(ROUND_ELT(sharedMemory, tmp, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+24]);
n[4] = xor1(ROUND_ELT(sharedMemory, tmp, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+24]);
n[5] = xor1(ROUND_ELT(sharedMemory, tmp, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+24]);
n[6] = xor1(ROUND_ELT(sharedMemory, tmp, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+24]);
n[7] = xor1(ROUND_ELT(sharedMemory, tmp, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+24]);
tmp[0] = xor1(ROUND_ELT(sharedMemory, n, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+32]);
tmp[1] = xor1(ROUND_ELT(sharedMemory, n, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+32]);
tmp[2] = xor1(ROUND_ELT(sharedMemory, n, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+32]);
tmp[3] = xor1(ROUND_ELT(sharedMemory, n, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+32]);
tmp[4] = xor1(ROUND_ELT(sharedMemory, n, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+32]);
tmp[5] = xor1(ROUND_ELT(sharedMemory, n, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+32]);
tmp[6] = xor1(ROUND_ELT(sharedMemory, n, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+32]);
tmp[7] = xor1(ROUND_ELT(sharedMemory, n, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+32]);
n[0] = xor1(ROUND_ELT(sharedMemory, tmp, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+40]);
n[1] = xor1(ROUND_ELT(sharedMemory, tmp, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+40]);
n[2] = xor1(ROUND_ELT(sharedMemory, tmp, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+40]);
n[3] = xor1(ROUND_ELT(sharedMemory, tmp, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+40]);
n[4] = xor1(ROUND_ELT(sharedMemory, tmp, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+40]);
n[5] = xor1(ROUND_ELT(sharedMemory, tmp, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+40]);
n[6] = xor1(ROUND_ELT(sharedMemory, tmp, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+40]);
n[7] = xor1(ROUND_ELT(sharedMemory, tmp, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+40]);
tmp[0] = xor1(ROUND_ELT(sharedMemory, n, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+48]);
tmp[1] = xor1(ROUND_ELT(sharedMemory, n, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+48]);
tmp[2] = xor1(ROUND_ELT(sharedMemory, n, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+48]);
tmp[3] = xor1(ROUND_ELT(sharedMemory, n, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+48]);
tmp[4] = xor1(ROUND_ELT(sharedMemory, n, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+48]);
tmp[5] = xor1(ROUND_ELT(sharedMemory, n, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+48]);
tmp[6] = xor1(ROUND_ELT(sharedMemory, n, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+48]);
tmp[7] = xor1(ROUND_ELT(sharedMemory, n, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+48]);
n[0] = xor1(ROUND_ELT(sharedMemory, tmp, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+56]);
n[1] = xor1(ROUND_ELT(sharedMemory, tmp, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+56]);
n[2] = xor1(ROUND_ELT(sharedMemory, tmp, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+56]);
n[3] = xor1(ROUND_ELT(sharedMemory, tmp, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+56]);
n[4] = xor1(ROUND_ELT(sharedMemory, tmp, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+56]);
n[5] = xor1(ROUND_ELT(sharedMemory, tmp, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+56]);
n[6] = xor1(ROUND_ELT(sharedMemory, tmp, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+56]);
n[7] = xor1(ROUND_ELT(sharedMemory, tmp, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+56]);
tmp[0] = xor1(ROUND_ELT(sharedMemory, n, 0, 7, 6, 5, 4, 3, 2, 1), c_tmp[0+64]);
tmp[1] = xor1(ROUND_ELT(sharedMemory, n, 1, 0, 7, 6, 5, 4, 3, 2), c_tmp[1+64]);
tmp[2] = xor1(ROUND_ELT(sharedMemory, n, 2, 1, 0, 7, 6, 5, 4, 3), c_tmp[2+64]);
tmp[3] = xor1(ROUND_ELT(sharedMemory, n, 3, 2, 1, 0, 7, 6, 5, 4), c_tmp[3+64]);
tmp[4] = xor1(ROUND_ELT(sharedMemory, n, 4, 3, 2, 1, 0, 7, 6, 5), c_tmp[4+64]);
tmp[5] = xor1(ROUND_ELT(sharedMemory, n, 5, 4, 3, 2, 1, 0, 7, 6), c_tmp[5+64]);
tmp[6] = xor1(ROUND_ELT(sharedMemory, n, 6, 5, 4, 3, 2, 1, 0, 7), c_tmp[6+64]);
tmp[7] = xor1(ROUND_ELT(sharedMemory, n, 7, 6, 5, 4, 3, 2, 1, 0), c_tmp[7+64]);
if (xor3(c_xtra[1], ROUND_ELT(sharedMemory, tmp, 3, 2, 1, 0, 7, 6, 5, 4), ROUND_ELT(sharedMemory, tmp, 5, 4, 3, 2, 1, 0, 7, 6)) <= pTarget[3]) {
atomicMin(&resNounce[0], nounce);
}
}
}
__host__
extern void whirlpoolx_cpu_init(int thr_id, uint32_t threads)
{
cudaMemcpyToSymbol(InitVector_RC, plain_RC, sizeof(plain_RC), 0, cudaMemcpyHostToDevice);
cudaMemcpyToSymbol(mixTob0Tox, plain_T0, sizeof(plain_T0), 0, cudaMemcpyHostToDevice);
cudaMalloc(&d_WXNonce[thr_id], sizeof(uint32_t));
cudaMallocHost(&h_wxnounce[thr_id], sizeof(uint32_t));
cudaMalloc(&d_xtra[thr_id], 8 * sizeof(uint64_t));
CUDA_SAFE_CALL(cudaMalloc(&d_tmp[thr_id], 8 * 9 * sizeof(uint64_t))); // d_tmp[threadIdx.x+64] (7+64)
}
__host__
void whirlpoolx_setBlock_80(void *pdata, const void *ptarget)
{
uint64_t PaddedMessage[16];
memcpy(PaddedMessage, pdata, 80);
memset((uint8_t*)&PaddedMessage+80, 0, 48);
((uint8_t*)PaddedMessage)[80] = 0x80; /* ending */
cudaMemcpyToSymbol(pTarget, ptarget, 4*sizeof(uint64_t), 0, cudaMemcpyHostToDevice);
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_PaddedMessage80, PaddedMessage, 16 * sizeof(uint64_t), 0, cudaMemcpyHostToDevice));
}
__host__
void whirlpoolx_precompute(int thr_id)
{
dim3 grid(1);
dim3 block(256);
whirlpoolx_gpu_precompute <<<grid, block>>>(8, d_xtra[thr_id], d_tmp[thr_id]);
cudaThreadSynchronize();
cudaMemcpyToSymbol(c_xtra, d_xtra[thr_id], 8 * sizeof(uint64_t), 0, cudaMemcpyDeviceToDevice);
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_tmp, d_tmp[thr_id], 8 * 9 * sizeof(uint64_t), 0, cudaMemcpyDeviceToDevice));
}
__host__
uint32_t whirlpoolx_cpu_hash(int thr_id, uint32_t threads, uint32_t startNounce)
{
dim3 grid((threads + threadsPerBlock-1) / threadsPerBlock);
dim3 block(threadsPerBlock);
cudaMemset(d_WXNonce[thr_id], 0xff, sizeof(uint32_t));
whirlpoolx_gpu_hash<<<grid, block>>>(threads, startNounce, d_WXNonce[thr_id]);
cudaThreadSynchronize();
cudaMemcpy(h_wxnounce[thr_id], d_WXNonce[thr_id], sizeof(uint32_t), cudaMemcpyDeviceToHost);
return *(h_wxnounce[thr_id]);
}
-97
View File
@@ -1,97 +0,0 @@
/*
* whirlpool routine (djm)
* whirlpoolx routine (provos alexis, tpruvot)
*/
extern "C" {
#include "sph/sph_whirlpool.h"
}
#include "miner.h"
#include "cuda_helper.h"
static uint32_t *d_hash[MAX_GPUS];
extern void whirlpoolx_cpu_init(int thr_id, uint32_t threads);
extern void whirlpoolx_setBlock_80(void *pdata, const void *ptarget);
extern uint32_t whirlpoolx_cpu_hash(int thr_id, uint32_t threads, uint32_t startNounce);
extern void whirlpoolx_precompute(int thr_id);
// CPU Hash function
extern "C" void whirlxHash(void *state, const void *input)
{
sph_whirlpool_context ctx_whirlpool;
unsigned char hash[64];
unsigned char hash_xored[32];
sph_whirlpool_init(&ctx_whirlpool);
sph_whirlpool(&ctx_whirlpool, input, 80);
sph_whirlpool_close(&ctx_whirlpool, hash);
// compress the 48 first bytes of the hash to 32
for (int i = 0; i < 32; i++) {
hash_xored[i] = hash[i] ^ hash[i + 16];
}
memcpy(state, hash_xored, 32);
}
static bool init[MAX_GPUS] = { 0 };
extern "C" int scanhash_whirlpoolx(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
{
const uint32_t first_nonce = pdata[19];
uint32_t endiandata[20];
int intensity = is_windows() ? 20 : 22;
uint32_t throughput = device_intensity(thr_id, __func__, 1U << intensity);
throughput = min(throughput, max_nonce - first_nonce);
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x0000ff;
if (!init[thr_id]) {
cudaSetDevice(device_map[thr_id]);
CUDA_CALL_OR_RET_X(cudaMalloc(&d_hash[thr_id], 64 * throughput), 0);
whirlpoolx_cpu_init(thr_id, throughput);
init[thr_id] = true;
}
for (int k=0; k < 20; k++) {
be32enc(&endiandata[k], pdata[k]);
}
whirlpoolx_setBlock_80((void*)endiandata, ptarget);
whirlpoolx_precompute(thr_id);
do {
uint32_t foundNonce = whirlpoolx_cpu_hash(thr_id, throughput, pdata[19]);
if (foundNonce != UINT32_MAX)
{
const uint32_t Htarg = ptarget[7];
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
whirlxHash(vhash64, endiandata);
if (vhash64[7] <= Htarg && fulltest(vhash64, ptarget)) {
*hashes_done = pdata[19] - first_nonce + throughput;
pdata[19] = foundNonce;
return 1;
} else {
applog(LOG_WARNING, "GPU #%d: result for %08x does not validate on CPU!", device_map[thr_id], foundNonce);
}
}
pdata[19] += throughput;
if (((uint64_t)pdata[19]+throughput) >= max_nonce) {
break;
}
} while (!work_restart[thr_id].restart);
*(hashes_done) = pdata[19] - first_nonce + 1;
return 0;
}
+2 -2
View File
@@ -49,7 +49,7 @@ extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -208,7 +208,7 @@ extern "C" int scanhash_x14(int thr_id, uint32_t *pdata,
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
+2 -2
View File
@@ -50,7 +50,7 @@ extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -219,7 +219,7 @@ extern "C" int scanhash_x15(int thr_id, uint32_t *pdata,
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
+2 -2
View File
@@ -53,7 +53,7 @@ extern void quark_jh512_cpu_init(int thr_id, uint32_t threads);
extern void quark_jh512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_init(int thr_id, uint32_t threads);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t *d_hash, int order);
extern void x11_luffaCubehash512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, uint32_t threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
@@ -243,7 +243,7 @@ extern "C" int scanhash_x17(int thr_id, uint32_t *pdata,
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, d_hash[thr_id], order++);
x11_luffaCubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
+12 -58
View File
@@ -33,9 +33,6 @@ static uint32_t* d_groes[MAX_GPUS];
static uint32_t* d_jh512[MAX_GPUS];
static uint32_t* d_skein[MAX_GPUS];
static uint8_t* d_txs[MAX_GPUS];
__constant__ uint16_t c_txlens[POK_MAX_TXS];
__constant__ uint8_t c_permut[24][4];
static const uint8_t permut[24][4] = {
{0, 1, 2, 3},
@@ -115,8 +112,8 @@ extern "C" void zr5hash(void *output, const void *input)
extern "C" void zr5hash_pok(void *output, uint32_t *pdata)
{
const uint32_t version = pdata[0] & (~POK_DATA_MASK);
uint32_t _ALIGN(64) hash[8];
const uint32_t version = (pdata[0] & (~POK_DATA_MASK)) | (use_pok ? POK_BOOL_MASK : 0);
pdata[0] = version;
zr5hash(hash, pdata);
@@ -211,7 +208,7 @@ void zr5_move_data_to_hash(int thr_id, uint32_t threads, int rnd)
__global__ __launch_bounds__(128, 8)
void zr5_get_poks_gpu(uint32_t threads, uint32_t** const d_buffers, uint8_t* const d_permut, uint16_t *d_poks)
void zr5_get_poks_gpu(uint32_t threads, uint32_t** const d_buffers, uint8_t* const d_permut, uint32_t *d_hash, uint16_t *d_poks)
{
const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
@@ -226,54 +223,14 @@ void zr5_get_poks_gpu(uint32_t threads, uint32_t** const d_buffers, uint8_t* con
}
}
__global__ __launch_bounds__(128, 4)
void zr5_get_poks_xor_gpu(uint32_t threads, uint32_t** const d_buffers, uint8_t* d_permut, uint16_t* d_poks, uint8_t* d_txs, uint8_t txs)
{
const uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
const uint8_t norder = d_permut[thread];
const uint8_t algo = c_permut[norder][3];
const uint8_t ntx = norder % txs; // generally 0 on testnet...
const uint32_t offset = thread * 16U; // 64 / 4;
uint32_t* hash = (uint32_t*) (d_buffers[algo] + offset);
uint32_t randNdx = hash[1] % c_txlens[ntx];
uint8_t* ptx = &d_txs[POK_MAX_TX_SZ*ntx] + randNdx;
uint32_t x = 0x100UL * ptx[3] + ptx[2];
d_poks[thread] = x ^ (hash[2] >> 16);
}
}
__host__
void zr5_get_poks(int thr_id, uint32_t threads, uint16_t* d_poks, struct work* work)
void zr5_get_poks(int thr_id, uint32_t threads, uint16_t* d_poks)
{
const uint32_t threadsperblock = 128;
dim3 grid((threads + threadsperblock - 1) / threadsperblock);
dim3 block(threadsperblock);
uint8_t txs = (uint8_t) work->tx_count;
if (txs && use_pok)
{
uint32_t txlens[POK_MAX_TXS];
uint8_t* txdata = (uint8_t*) calloc(POK_MAX_TXS, POK_MAX_TX_SZ);
if (!txdata) {
applog(LOG_ERR, "%s: error, memory alloc failure", __func__);
return;
}
// create blocs to copy on device
for (uint8_t tx=0; tx < txs; tx++) {
txlens[tx] = (uint32_t) (work->txs[tx].len - 3U);
memcpy(&txdata[POK_MAX_TX_SZ*tx], work->txs[tx].data, min(POK_MAX_TX_SZ, txlens[tx]+3U));
}
cudaMemcpy(d_txs[thr_id], txdata, txs * POK_MAX_TX_SZ, cudaMemcpyHostToDevice);
CUDA_SAFE_CALL(cudaMemcpyToSymbol(c_txlens, txlens, txs * sizeof(uint32_t), 0, cudaMemcpyHostToDevice));
zr5_get_poks_xor_gpu <<<grid, block>>> (threads, d_buffers[thr_id], d_permut[thr_id], d_poks, d_txs[thr_id], txs);
free(txdata);
} else {
zr5_get_poks_gpu <<<grid, block>>> (threads, d_buffers[thr_id], d_permut[thr_id], d_poks);
}
zr5_get_poks_gpu <<<grid, block>>> (threads, d_buffers[thr_id], d_permut[thr_id], d_hash[thr_id], d_poks);
}
@@ -328,21 +285,18 @@ extern void quark_skein512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t st
static bool init[MAX_GPUS] = { 0 };
extern "C" int scanhash_zr5(int thr_id, struct work *work,
extern "C" int scanhash_zr5(int thr_id, uint32_t *pdata, const uint32_t *ptarget,
uint32_t max_nonce, unsigned long *hashes_done)
{
uint32_t _ALIGN(64) tmpdata[20];
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t oldp0 = pdata[0];
const uint32_t version = (oldp0 & (~POK_DATA_MASK)) | (use_pok ? POK_BOOL_MASK : 0);
const uint32_t version = pdata[0] & (~POK_DATA_MASK);
const uint32_t first_nonce = pdata[19];
uint32_t throughput = device_intensity(thr_id, __func__, 1U << 18);
throughput = min(throughput, (1U << 20)-1024);
throughput = min(throughput, max_nonce - first_nonce);
if (opt_benchmark)
ptarget[7] = 0x0000ff;
((uint32_t*)ptarget)[7] = 0x0000ff;
memcpy(tmpdata, pdata, 80);
@@ -365,8 +319,6 @@ extern "C" int scanhash_zr5(int thr_id, struct work *work,
cudaMalloc(&d_jh512[thr_id], 64 * throughput);
cudaMalloc(&d_skein[thr_id], 64 * throughput);
cudaMalloc(&d_txs[thr_id], POK_MAX_TXS * POK_MAX_TX_SZ);
jackpot_keccak512_cpu_init(thr_id, throughput);
quark_blake512_cpu_init(thr_id, throughput);
@@ -402,7 +354,7 @@ extern "C" int scanhash_zr5(int thr_id, struct work *work,
}
// store on device d_poks all hash[0] prefixes
zr5_get_poks(thr_id, throughput, d_poks[thr_id], work);
zr5_get_poks(thr_id, throughput, d_poks[thr_id]);
// Keccak512 with pok
zr5_keccak512_cpu_hash_pok(thr_id, throughput, pdata[19], pdata, d_hash[thr_id], d_poks[thr_id]);
@@ -422,6 +374,7 @@ extern "C" int scanhash_zr5(int thr_id, struct work *work,
if (foundNonce != UINT32_MAX)
{
uint32_t vhash64[8];
uint32_t oldp0 = pdata[0];
uint32_t oldp19 = pdata[19];
uint32_t offset = foundNonce - pdata[19];
uint32_t pok = 0;
@@ -453,6 +406,9 @@ extern "C" int scanhash_zr5(int thr_id, struct work *work,
} else {
applog(LOG_WARNING, "GPU #%d: result for %08x does not validate on CPU!", device_map[thr_id], foundNonce);
// reinit the card.. segfault so no
// cuda_reset_device(thr_id, init);
pdata[19]++;
pdata[0] = oldp0;
}
@@ -461,8 +417,6 @@ extern "C" int scanhash_zr5(int thr_id, struct work *work,
} while (pdata[19] < max_nonce && !work_restart[thr_id].restart);
pdata[0] = oldp0;
*hashes_done = pdata[19] - first_nonce + 1;
return 0;
}