GOSTCoin CUDA miner project, compatible with most nvidia cards, containing only gostd algo
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

279 lines
10 KiB

/*
* X13 algorithm
*/
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 "sph/sph_hamsi.h"
#include "sph/sph_fugue.h"
#include "miner.h"
}
#include <stdint.h>
#include <cuda_runtime.h>
// aus cpu-miner.c
extern int device_map[8];
// Speicher f<EFBFBD>r Input/Output der verketteten Hashfunktionen
static uint32_t *d_hash[8];
extern void quark_blake512_cpu_init(int thr_id, int threads);
extern void quark_blake512_cpu_setBlock_80(void *pdata);
extern void quark_blake512_cpu_hash_80(int thr_id, int threads, uint32_t startNounce, uint32_t *d_hash, int order);
extern void quark_blake512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_bmw512_cpu_init(int thr_id, int threads);
extern void quark_bmw512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_groestl512_cpu_init(int thr_id, int threads);
extern void quark_groestl512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
//extern void quark_doublegroestl512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_skein512_cpu_init(int thr_id, int threads);
extern void quark_skein512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_keccak512_cpu_init(int thr_id, int threads);
extern void quark_keccak512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_jh512_cpu_init(int thr_id, int threads);
extern void quark_jh512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_luffa512_cpu_init(int thr_id, int threads);
extern void x11_luffa512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_cubehash512_cpu_init(int thr_id, int threads);
extern void x11_cubehash512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_shavite512_cpu_init(int thr_id, int threads);
extern void x11_shavite512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_simd512_cpu_init(int thr_id, int threads);
extern void x11_simd512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x11_echo512_cpu_init(int thr_id, int threads);
extern void x11_echo512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x13_hamsi512_cpu_init(int thr_id, int threads);
extern void x13_hamsi512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void x13_fugue512_cpu_init(int thr_id, int threads);
extern void x13_fugue512_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_hash, int order);
extern void quark_check_cpu_init(int thr_id, int threads);
extern void quark_check_cpu_setTarget(const void *ptarget);
extern uint32_t quark_check_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint32_t *d_inputHash, int order);
extern void quark_compactTest_cpu_init(int thr_id, int threads);
extern void quark_compactTest_cpu_hash_64(int thr_id, int threads, uint32_t startNounce, uint32_t *inpHashes,
uint32_t *d_noncesTrue, size_t *nrmTrue, uint32_t *d_noncesFalse, size_t *nrmFalse,
int order);
// X13 Hashfunktion
extern "C" void x13hash(void *output, const void *input)
{
// blake1-bmw2-grs3-skein4-jh5-keccak6-luffa7-cubehash8-shavite9-simd10-echo11-hamsi12-fugue13
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;
sph_hamsi512_context ctx_hamsi;
sph_fugue512_context ctx_fugue;
uint32_t hash[32];
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_skein512_init(&ctx_skein);
sph_skein512 (&ctx_skein, (const void*) hash, 64);
sph_skein512_close(&ctx_skein, (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_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);
sph_hamsi512_init(&ctx_hamsi);
sph_hamsi512 (&ctx_hamsi, (const void*) hash, 64);
sph_hamsi512_close(&ctx_hamsi, (void*) hash);
sph_fugue512_init(&ctx_fugue);
sph_fugue512 (&ctx_fugue, (const void*) hash, 64);
sph_fugue512_close(&ctx_fugue, (void*) hash);
memcpy(output, hash, 32);
}
extern bool opt_benchmark;
extern "C" int scanhash_x13(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];
if (opt_benchmark)
((uint32_t*)ptarget)[7] = 0x0000ff;
const uint32_t Htarg = ptarget[7];
const int throughput = 256*256*8;
static bool init[8] = {0,0,0,0,0,0,0,0};
if (!init[thr_id])
{
cudaSetDevice(device_map[thr_id]);
// Konstanten kopieren, Speicher belegen
cudaMalloc(&d_hash[thr_id], 16 * sizeof(uint32_t) * throughput);
quark_blake512_cpu_init(thr_id, throughput);
quark_groestl512_cpu_init(thr_id, throughput);
quark_skein512_cpu_init(thr_id, throughput);
quark_bmw512_cpu_init(thr_id, throughput);
quark_keccak512_cpu_init(thr_id, throughput);
quark_jh512_cpu_init(thr_id, throughput);
x11_luffa512_cpu_init(thr_id, throughput);
x11_cubehash512_cpu_init(thr_id, throughput);
x11_shavite512_cpu_init(thr_id, throughput);
x11_simd512_cpu_init(thr_id, throughput);
x11_echo512_cpu_init(thr_id, throughput);
x13_hamsi512_cpu_init(thr_id, throughput);
x13_fugue512_cpu_init(thr_id, throughput);
quark_check_cpu_init(thr_id, throughput);
init[thr_id] = true;
}
//unsigned char echobefore[64], echoafter[64];
uint32_t endiandata[20];
for (int k=0; k < 20; k++)
be32enc(&endiandata[k], ((uint32_t*)pdata)[k]);
quark_blake512_cpu_setBlock_80((void*)endiandata);
quark_check_cpu_setTarget(ptarget);
do {
int order = 0;
// erstes Blake512 Hash mit CUDA
quark_blake512_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r BMW512
quark_bmw512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Groestl512
quark_groestl512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Skein512
quark_skein512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r JH512
quark_jh512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Keccak512
quark_keccak512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Luffa512
x11_luffa512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Cubehash512
x11_cubehash512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r Shavite512
x11_shavite512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r SIMD512
x11_simd512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// das ist der unbedingte Branch f<EFBFBD>r ECHO512
x11_echo512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x13_hamsi512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
x13_fugue512_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
// Scan nach Gewinner Hashes auf der GPU
uint32_t foundNonce = quark_check_cpu_hash_64(thr_id, throughput, pdata[19], NULL, d_hash[thr_id], order++);
if (foundNonce != 0xffffffff)
{
uint32_t vhash64[8];
be32enc(&endiandata[19], foundNonce);
x13hash(vhash64, endiandata);
if( (vhash64[7]<=Htarg) && fulltest(vhash64, ptarget) ) {
pdata[19] = foundNonce;
*hashes_done = foundNonce - first_nonce + 1;
return 1;
}
else if (vhash64[7] > Htarg) {
applog(LOG_INFO, "GPU #%d: result for %08x is not in range: %x > %x", thr_id, foundNonce, vhash64[7], Htarg);
}
else {
applog(LOG_INFO, "GPU #%d: result for %08x does not validate on CPU!", thr_id, foundNonce);
}
}
pdata[19] += throughput;
} while (pdata[19] < max_nonce && !work_restart[thr_id].restart);
*hashes_done = pdata[19] - first_nonce + 1;
return 0;
}