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.
134 lines
4.5 KiB
134 lines
4.5 KiB
![]()
10 years ago
|
extern "C" {
|
||
|
#include "sph/sph_blake.h"
|
||
|
#include "sph/sph_groestl.h"
|
||
|
#include "sph/sph_skein.h"
|
||
|
#include "sph/sph_keccak.h"
|
||
|
#include "lyra2/Lyra2.h"
|
||
|
}
|
||
|
|
||
|
#include "miner.h"
|
||
|
#include "cuda_helper.h"
|
||
|
|
||
|
static _ALIGN(64) uint64_t *d_hash[8];
|
||
|
|
||
|
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 uint32_t quark_check_cpu_hash_64_2(int thr_id, int threads, uint32_t startNounce, uint32_t *d_nonceVector, uint64_t *d_inputHash, int order);
|
||
|
|
||
|
extern void blake256_cpu_init(int thr_id, int threads);
|
||
|
extern void blake256_cpu_hash_80(const int thr_id, const uint32_t threads, const uint32_t startNonce, uint64_t *Hash, int order);
|
||
|
extern void blake256_cpu_setBlock_80(uint32_t *pdata);
|
||
|
extern void keccak256_cpu_hash_32(int thr_id, int threads, uint32_t startNonce, uint64_t *d_outputHash, int order);
|
||
|
extern void keccak256_cpu_init(int thr_id, int threads);
|
||
|
extern void skein256_cpu_hash_32(int thr_id, int threads, uint32_t startNonce, uint64_t *d_outputHash, int order);
|
||
|
extern void skein256_cpu_init(int thr_id, int threads);
|
||
|
|
||
|
extern void lyra2_cpu_hash_32(int thr_id, int threads, uint32_t startNonce, uint64_t *d_outputHash, int order);
|
||
|
extern void lyra2_cpu_init(int thr_id, int threads);
|
||
|
|
||
|
extern void groestl256_setTarget(const void *ptarget);
|
||
|
extern uint32_t groestl256_cpu_hash_32(int thr_id, int threads, uint32_t startNounce, uint64_t *d_outputHash, int order);
|
||
|
extern void groestl256_cpu_init(int thr_id, int threads);
|
||
|
|
||
|
extern "C" void lyra_hash(void *state, const void *input)
|
||
|
{
|
||
|
sph_blake256_context ctx_blake;
|
||
|
sph_keccak256_context ctx_keccak;
|
||
|
sph_skein256_context ctx_skein;
|
||
|
sph_groestl256_context ctx_groestl;
|
||
|
|
||
|
uint32_t hashA[8], hashB[8], hash[8];
|
||
|
|
||
|
sph_blake256_init(&ctx_blake);
|
||
|
sph_blake256(&ctx_blake, input, 80);
|
||
|
sph_blake256_close(&ctx_blake, hashA);
|
||
|
|
||
|
sph_keccak256_init(&ctx_keccak);
|
||
|
sph_keccak256(&ctx_keccak, hashA, 32);
|
||
|
sph_keccak256_close(&ctx_keccak, hashB);
|
||
|
|
||
|
LYRA2(hashA, 32, hashB, 32, hashB, 32, 1, 8, 8);
|
||
|
|
||
|
sph_skein256_init(&ctx_skein);
|
||
|
sph_skein256(&ctx_skein, hashA, 32);
|
||
|
sph_skein256_close(&ctx_skein, hashB);
|
||
|
|
||
|
sph_groestl256_init(&ctx_groestl);
|
||
|
sph_groestl256(&ctx_groestl, hashB, 32);
|
||
|
sph_groestl256_close(&ctx_groestl, hash);
|
||
|
|
||
|
// seems wrong : hash or hashB ?
|
||
|
memcpy(state, hashB, 32);
|
||
|
}
|
||
|
|
||
|
static bool init[8] = { 0 };
|
||
|
|
||
|
extern "C" int scanhash_lyra(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) ? 19 : 18;
|
||
|
int throughput = opt_work_size ? opt_work_size : (1 << intensity); // 18=256*256*4;
|
||
|
throughput = min(throughput, (int)(max_nonce - first_nonce));
|
||
|
|
||
|
if (opt_benchmark)
|
||
|
((uint32_t*)ptarget)[7] = 0x0000ff;
|
||
|
|
||
|
if (!init[thr_id])
|
||
|
{
|
||
|
cudaSetDevice(device_map[thr_id]);
|
||
|
|
||
|
blake256_cpu_init(thr_id, throughput);
|
||
|
keccak256_cpu_init(thr_id,throughput);
|
||
|
skein256_cpu_init(thr_id, throughput);
|
||
|
groestl256_cpu_init(thr_id, throughput);
|
||
|
lyra2_cpu_init(thr_id, throughput);
|
||
|
|
||
|
CUDA_SAFE_CALL(cudaMalloc(&d_hash[thr_id], 16 * sizeof(uint32_t) * throughput));
|
||
|
|
||
|
init[thr_id] = true;
|
||
|
}
|
||
|
|
||
|
uint32_t endiandata[20];
|
||
|
for (int k=0; k < 20; k++)
|
||
|
be32enc(&endiandata[k], ((uint32_t*)pdata)[k]);
|
||
|
|
||
|
blake256_cpu_setBlock_80(pdata);
|
||
|
groestl256_setTarget(ptarget);
|
||
|
|
||
|
do {
|
||
|
int order = 0;
|
||
|
uint32_t foundNonce;
|
||
|
|
||
|
blake256_cpu_hash_80(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
|
||
|
keccak256_cpu_hash_32(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
|
||
|
lyra2_cpu_hash_32(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
|
||
|
skein256_cpu_hash_32(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
|
||
|
|
||
|
foundNonce = groestl256_cpu_hash_32(thr_id, throughput, pdata[19], d_hash[thr_id], order++);
|
||
|
if (foundNonce != 0xffffffff)
|
||
|
{
|
||
|
// const uint32_t Htarg = ptarget[6];
|
||
|
uint32_t vhash64[8];
|
||
|
be32enc(&endiandata[19], foundNonce);
|
||
|
lyra_hash(vhash64, endiandata);
|
||
|
|
||
|
// if (vhash64[7]<=Htarg) { // && fulltest(vhash64, ptarget)) {
|
||
|
*hashes_done = pdata[19] - first_nonce + throughput;
|
||
|
pdata[19] = foundNonce;
|
||
|
return 1;
|
||
|
// } else {
|
||
|
// applog(LOG_INFO, "GPU #%d: result for nonce $%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;
|
||
|
}
|