Browse Source

blake80: some changes and launch bounds, no perf changes

master
Tanguy Pruvot 10 years ago
parent
commit
2113be6eec
  1. 8
      cuda_nist5.cu
  2. 92
      quark/cuda_quark_blake512.cu
  3. 8
      quark/quarkcoin.cu
  4. 9
      x11/x11.cu
  5. 9
      x13/x13.cu
  6. 8
      x15/x14.cu
  7. 8
      x15/x15.cu
  8. 8
      x17/x17.cu

8
cuda_nist5.cu

@ -14,8 +14,8 @@ extern "C" @@ -14,8 +14,8 @@ extern "C"
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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_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);
@ -99,14 +99,14 @@ extern "C" int scanhash_nist5(int thr_id, uint32_t *pdata, @@ -99,14 +99,14 @@ extern "C" int scanhash_nist5(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
int order = 0;
// Hash with CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
quark_groestl512_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++);

92
quark/cuda_quark_blake512.cu

@ -7,14 +7,13 @@ @@ -7,14 +7,13 @@
#define USE_SHUFFLE 0
__constant__ uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
__constant__
static uint64_t c_PaddedMessage80[16]; // padded message (80 bytes + padding)
// ---------------------------- BEGIN CUDA quark_blake512 functions ------------------------------------
__constant__ uint8_t c_sigma[16][16];
const uint8_t host_sigma[16][16] =
{
__device__ __constant__
static const uint8_t c_sigma[16][16] = {
{ 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 },
{14, 10, 4, 8, 9, 15, 13, 6, 1, 12, 0, 2, 11, 7, 5, 3 },
{11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4 },
@ -30,11 +29,10 @@ const uint8_t host_sigma[16][16] = @@ -30,11 +29,10 @@ const uint8_t host_sigma[16][16] =
{11, 8, 12, 0, 5, 2, 15, 13, 10, 14, 3, 6, 7, 1, 9, 4 },
{ 7, 9, 3, 1, 13, 12, 11, 14, 2, 6, 5, 10, 4, 0, 15, 8 },
{ 9, 0, 5, 7, 2, 4, 10, 15, 14, 1, 11, 12, 6, 8, 3, 13 },
{ 2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9 }
};
{ 2, 12, 6, 10, 0, 11, 8, 3, 4, 13, 7, 5, 15, 14, 1, 9 }};
__device__ __constant__
const uint64_t c_u512[16] =
static const uint64_t c_u512[16] =
{
0x243f6a8885a308d3ULL, 0x13198a2e03707344ULL,
0xa4093822299f31d0ULL, 0x082efa98ec4e6c89ULL,
@ -111,14 +109,8 @@ void quark_blake512_compress( uint64_t *h, const uint64_t *block, const uint8_t @@ -111,14 +109,8 @@ void quark_blake512_compress( uint64_t *h, const uint64_t *block, const uint8_t
// Hash-Padding
__device__ __constant__
static const uint64_t d_constHashPadding[8] = {
0x0000000000000080ull,
0,
0,
0,
0,
0x0100000000000000ull,
0,
0x0002000000000000ull
0x0000000000000080ull, 0, 0, 0,
0, 0x0100000000000000ull, 0, 0x0002000000000000ull
};
__global__ __launch_bounds__(256, 4)
@ -171,14 +163,14 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t @@ -171,14 +163,14 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t
quark_blake512_compress( h, buf, c_sigma, c_u512, 512 );
#if __CUDA_ARCH__ <= 350
uint32_t *outHash = (uint32_t*)&g_hash[8 * hashPosition];
uint32_t *outHash = (uint32_t*)&g_hash[hashPosition * 8U];
#pragma unroll 8
for (int i=0; i < 8; i++) {
outHash[2*i+0] = cuda_swab32( _HIWORD(h[i]) );
outHash[2*i+1] = cuda_swab32( _LOWORD(h[i]) );
}
#else
uint64_t *outHash = &g_hash[8 * hashPosition];
uint64_t *outHash = &g_hash[hashPosition * 8U];
for (int i=0; i < 8; i++) {
outHash[i] = cuda_swab64(h[i]);
}
@ -186,13 +178,20 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t @@ -186,13 +178,20 @@ void quark_blake512_gpu_hash_64(uint32_t threads, uint32_t startNounce, uint32_t
}
}
__global__ void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounce, void *outputHash)
__global__ __launch_bounds__(256,4)
void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounce, void *outputHash)
{
uint32_t thread = (blockDim.x * blockIdx.x + threadIdx.x);
if (thread < threads)
{
uint64_t buf[16];
uint32_t nounce = startNounce + thread;
#pragma unroll
for (int i=0; i < 16; ++i)
buf[i] = c_PaddedMessage80[i];
// The test Nonce
const uint32_t nounce = startNounce + thread;
((uint32_t*)buf)[19] = cuda_swab32(nounce);
uint64_t h[8] = {
0x6a09e667f3bcc908ULL,
@ -205,30 +204,21 @@ __global__ void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounc @@ -205,30 +204,21 @@ __global__ void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounc
0x5be0cd19137e2179ULL
};
// Message für die erste Runde in Register holen
#pragma unroll 16
for (int i=0; i < 16; ++i)
buf[i] = c_PaddedMessage80[i];
// The test Nonce
((uint32_t*)buf)[19] = cuda_swab32(nounce);
quark_blake512_compress( h, buf, c_sigma, c_u512, 640 );
quark_blake512_compress(h, buf, c_sigma, c_u512, 640);
#if __CUDA_ARCH__ <= 350
uint32_t *outHash = (uint32_t *)outputHash + 16 * thread;
uint32_t *outHash = (uint32_t*)outputHash + (thread * 16U);
#pragma unroll 8
for (uint32_t i=0; i < 8; i++) {
outHash[2*i] = cuda_swab32( _HIWORD(h[i]) );
outHash[2*i+1] = cuda_swab32( _LOWORD(h[i]) );
}
#else
uint64_t *outHash = (uint64_t *)outputHash + 8 * thread;
uint64_t *outHash = (uint64_t*)outputHash + (thread * 8U);
for (uint32_t i=0; i < 8; i++) {
outHash[i] = cuda_swab64( h[i] );
}
#endif
}
}
@ -238,30 +228,24 @@ __global__ void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounc @@ -238,30 +228,24 @@ __global__ void quark_blake512_gpu_hash_80(uint32_t threads, uint32_t startNounc
__host__
void quark_blake512_cpu_init(int thr_id, uint32_t threads)
{
// Kopiere die Hash-Tabellen in den GPU-Speicher
CUDA_CALL_OR_RET( cudaMemcpyToSymbol(c_sigma,
host_sigma,
sizeof(host_sigma),
0, cudaMemcpyHostToDevice));
CUDA_SAFE_CALL(cudaGetLastError());
}
// Blake512 für 80 Byte grosse Eingangsdaten
__host__
void quark_blake512_cpu_setBlock_80(void *pdata)
void quark_blake512_cpu_setBlock_80(int thr_id, uint32_t *pdata)
{
// Message mit Padding bereitstellen
// lediglich die korrekte Nonce ist noch ab Byte 76 einzusetzen.
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_PaddedMessage80, PaddedMessage, 16*sizeof(uint64_t), 0, cudaMemcpyHostToDevice)
);
uint64_t message[16];
memcpy(message, pdata, 80);
message[10] = 0x80;
message[11] = 0;
message[12] = 0;
message[13] = 0x0100000000000000ull;
message[14] = 0;
message[15] = 0x8002000000000000ull; // 0x280
cudaMemcpyToSymbol(c_PaddedMessage80, message, sizeof(message), 0, cudaMemcpyHostToDevice);
CUDA_SAFE_CALL(cudaGetLastError());
}
__host__
@ -269,18 +253,16 @@ void quark_blake512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNoun @@ -269,18 +253,16 @@ void quark_blake512_cpu_hash_64(int thr_id, uint32_t threads, uint32_t startNoun
{
const uint32_t threadsperblock = 256;
// berechne wie viele Thread Blocks wir brauchen
dim3 grid((threads + threadsperblock-1)/threadsperblock);
dim3 block(threadsperblock);
quark_blake512_gpu_hash_64<<<grid, block>>>(threads, startNounce, d_nonceVector, (uint64_t*)d_outputHash);
// Strategisches Sleep Kommando zur Senkung der CPU Last
//MyStreamSynchronize(NULL, order, thr_id);
}
__host__
void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_outputHash, int order)
void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_outputHash)
{
const uint32_t threadsperblock = 256;

8
quark/quarkcoin.cu

@ -20,8 +20,8 @@ static uint32_t *d_branch2Nonces[MAX_GPUS]; @@ -20,8 +20,8 @@ static uint32_t *d_branch2Nonces[MAX_GPUS];
static uint32_t *d_branch3Nonces[MAX_GPUS];
extern void quark_blake512_cpu_init(int thr_id, uint32_t threads);
extern void quark_blake512_cpu_setBlock_80(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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_blake512_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_bmw512_cpu_init(int thr_id, uint32_t threads);
@ -170,7 +170,7 @@ extern "C" int scanhash_quark(int thr_id, uint32_t *pdata, @@ -170,7 +170,7 @@ extern "C" int scanhash_quark(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
@ -178,7 +178,7 @@ extern "C" int scanhash_quark(int thr_id, uint32_t *pdata, @@ -178,7 +178,7 @@ extern "C" int scanhash_quark(int thr_id, uint32_t *pdata,
size_t nrm1=0, nrm2=0, nrm3=0;
// erstes Blake512 Hash mit CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
// das ist der unbedingte Branch für BMW512
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);

9
x11/x11.cu

@ -23,8 +23,8 @@ extern "C" @@ -23,8 +23,8 @@ extern "C"
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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);
@ -181,8 +181,7 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata, @@ -181,8 +181,7 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
@ -190,7 +189,7 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata, @@ -190,7 +189,7 @@ extern "C" int scanhash_x11(int thr_id, uint32_t *pdata,
uint32_t foundNonce;
// Hash with CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
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 :");

9
x13/x13.cu

@ -25,10 +25,9 @@ extern "C" @@ -25,10 +25,9 @@ extern "C"
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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);
@ -188,14 +187,14 @@ extern "C" int scanhash_x13(int thr_id, uint32_t *pdata, @@ -188,14 +187,14 @@ extern "C" int scanhash_x13(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
uint32_t foundNonce;
int order = 0;
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);

8
x15/x14.cu

@ -30,8 +30,8 @@ extern "C" { @@ -30,8 +30,8 @@ extern "C" {
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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);
@ -197,12 +197,12 @@ extern "C" int scanhash_x14(int thr_id, uint32_t *pdata, @@ -197,12 +197,12 @@ extern "C" int scanhash_x14(int thr_id, uint32_t *pdata,
for (int k = 0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
int order = 0;
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);

8
x15/x15.cu

@ -31,8 +31,8 @@ extern "C" { @@ -31,8 +31,8 @@ extern "C" {
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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);
@ -208,12 +208,12 @@ extern "C" int scanhash_x15(int thr_id, uint32_t *pdata, @@ -208,12 +208,12 @@ extern "C" int scanhash_x15(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
int order = 0;
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);

8
x17/x17.cu

@ -34,8 +34,8 @@ extern "C" @@ -34,8 +34,8 @@ extern "C"
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(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, uint32_t threads, uint32_t startNounce, uint32_t *d_hash, int order);
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);
@ -230,14 +230,14 @@ extern "C" int scanhash_x17(int thr_id, uint32_t *pdata, @@ -230,14 +230,14 @@ extern "C" int scanhash_x17(int thr_id, uint32_t *pdata,
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], pdata[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_blake512_cpu_setBlock_80(thr_id, endiandata);
cuda_check_cpu_setTarget(ptarget);
do {
int order = 0;
// Hash with CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id]); order++;
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);

Loading…
Cancel
Save