Files
cpuminer-opt-gpu/algo/quark/quark.c
Jay D Dee 7b94436202 v3.5.9.1
2017-03-04 16:23:24 -05:00

300 lines
7.7 KiB
C

#include "cpuminer-config.h"
#include "miner.h"
#include "algo-gate-api.h"
#include <stdio.h>
#include <string.h>
#include <stdint.h>
#include "algo/blake/sph_blake.h"
#include "algo/bmw/sph_bmw.h"
#include "algo/groestl/sph_groestl.h"
#include "algo/jh/sph_jh.h"
#include "algo/keccak/sph_keccak.h"
#include "algo/skein/sph_skein.h"
#include "algo/blake/sse2/blake.c"
#include "algo/bmw/sse2/bmw.c"
#include "algo/keccak/sse2/keccak.c"
#include "algo/skein/sse2/skein.c"
#include "algo/jh/sse2/jh_sse2_opt64.h"
#ifdef NO_AES_NI
#include "algo/groestl/sse2/grso.h"
#include "algo/groestl/sse2/grso-macro.c"
#else
#include "algo/groestl/aes_ni/hash-groestl.h"
#endif
/*define data alignment for different C compilers*/
#if defined(__GNUC__)
#define DATA_ALIGN16(x) x __attribute__ ((aligned(16)))
#define DATA_ALIGNXY(x,y) x __attribute__ ((aligned(y)))
#else
#define DATA_ALIGN16(x) __declspec(align(16)) x
#define DATA_ALIGNXY(x,y) __declspec(align(y)) x
#endif
#ifdef NO_AES_NI
sph_groestl512_context quark_ctx;
#else
hashState_groestl quark_ctx;
#endif
void init_quark_ctx()
{
#ifdef NO_AES_NI
sph_groestl512_init( &quark_ctx );
#else
init_groestl( &quark_ctx, 64 );
#endif
}
inline static void quarkhash(void *state, const void *input)
{
unsigned char hashbuf[128];
size_t hashptr;
sph_u64 hashctA;
sph_u64 hashctB;
int i;
unsigned char hash[128] __attribute__ ((aligned (32)));
#ifdef NO_AES_NI
sph_groestl512_context ctx;
#else
hashState_groestl ctx;
#endif
memcpy( &ctx, &quark_ctx, sizeof(ctx) );
// Blake
DECL_BLK;
BLK_I;
BLK_W;
for(i=0; i<9; i++)
{
/* blake is split between 64byte hashes and the 80byte initial block */
//DECL_BLK;
switch (i+(16*((hash[0] & (uint32_t)(8)) == (uint32_t)(0))))
{
// Blake
case 5 :
BLK_I;
BLK_U;
case 0:
case 16:
BLK_C;
break;
case 1:
case 17:
case 21:
// BMW
do
{
DECL_BMW;
BMW_I;
BMW_U;
/* bmw compress uses some defines */
/* i havent gotten around to rewriting these */
#define M(x) sph_dec64le_aligned(data + 8 * (x))
#define H(x) (h[x])
#define dH(x) (dh[x])
BMW_C;
#undef M
#undef H
#undef dH
} while(0); continue;;
case 2:
// dos this entry point represent a second groestl round?
case 3:
case 19:
// Groestl
do
{
#ifdef NO_AES_NI
grsoState sts_grs;
GRS_I;
GRS_U;
GRS_C;
// sph_groestl512_init( &ctx );
// sph_groestl512 ( &ctx, hash, 64 );
// sph_groestl512_close( &ctx, hash );
#else
reinit_groestl( &ctx );
update_groestl( &ctx, (char*)hash, 512 );
final_groestl( &ctx, (char*)hash );
#endif
} while(0); continue;
case 4:
case 20:
case 24:
// JH
do
{
DECL_JH;
JH_H;
} while(0); continue;
case 6:
case 22:
case 8:
// Keccak
do
{
DECL_KEC;
KEC_I;
KEC_U;
KEC_C;
} while(0); continue;
case 18:
case 7:
case 23:
// Skein
do
{
DECL_SKN;
SKN_I;
SKN_U;
SKN_C; /* is a magintue faster than others, done */
} while(0); continue;
default:
/* bad things happend, i counted to potato */
abort();
}
/* only blake shouuld get here without continue */
/* blake finishs from top split */
//BLK_C;
}
// asm volatile ("emms");
memcpy(state, hash, 32);
}
void quarkhash_alt(void *state, const void *input)
{
sph_blake512_context ctx_blake1,
ctx_blake2;
sph_bmw512_context ctx_bmw1,
ctx_bmw2;
sph_groestl512_context ctx_groestl1,
ctx_groestl2;
sph_skein512_context ctx_skein1,
ctx_skein2;
sph_jh512_context ctx_jh1,
ctx_jh2;
sph_keccak512_context ctx_keccak1,
ctx_keccak2;
sph_blake512_init(&ctx_blake1);
sph_bmw512_init(&ctx_bmw1);
sph_groestl512_init(&ctx_groestl1);
sph_skein512_init(&ctx_skein1);
sph_groestl512_init(&ctx_groestl2);
sph_jh512_init(&ctx_jh1);
sph_blake512_init(&ctx_blake2);
sph_bmw512_init(&ctx_bmw2);
sph_keccak512_init(&ctx_keccak1);
sph_skein512_init(&ctx_skein2);
sph_keccak512_init(&ctx_keccak2);
sph_jh512_init(&ctx_jh2);
uint32_t _ALIGN(128) hash[16];
uint32_t mask = 8;
sph_blake512 (&ctx_blake1, input, 80);
sph_blake512_close (&ctx_blake1, hash); //0
sph_bmw512 (&ctx_bmw1, hash, 64);
sph_bmw512_close(&ctx_bmw1, hash); //1
if (hash[0] & mask) {
sph_groestl512 (&ctx_groestl1, hash, 64);
sph_groestl512_close(&ctx_groestl1, hash); //2
} else {
sph_skein512 (&ctx_skein1, hash, 64);
sph_skein512_close(&ctx_skein1, hash); //2
}
sph_groestl512 (&ctx_groestl2, hash, 64);
sph_groestl512_close(&ctx_groestl2, hash); //3
sph_jh512 (&ctx_jh1, hash, 64);
sph_jh512_close(&ctx_jh1, hash); //4
if (hash[0] & mask) {
sph_blake512 (&ctx_blake2, hash, 64);
sph_blake512_close(&ctx_blake2, hash); //5
} else {
sph_bmw512 (&ctx_bmw2, hash, 64);
sph_bmw512_close(&ctx_bmw2, hash); //5
}
sph_keccak512 (&ctx_keccak1, hash, 64);
sph_keccak512_close(&ctx_keccak1, hash); //6
sph_skein512 (&ctx_skein2, hash, 64);
sph_skein512_close(&ctx_skein2, hash); //7
if (hash[0] & mask) {
sph_keccak512 (&ctx_keccak2, hash, 64);
sph_keccak512_close(&ctx_keccak2, hash); //8
} else {
sph_jh512 (&ctx_jh2, hash, 64);
sph_jh512_close(&ctx_jh2, hash); //8
}
memcpy(state, hash, 32);
}
int scanhash_quark( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done)
{
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
swab32_array( endiandata, pdata, 20 );
do {
pdata[19] = ++n;
be32enc(&endiandata[19], n);
quarkhash(hash64, &endiandata);
if ((hash64[7]&0xFFFFFF00)==0)
{
if (fulltest(hash64, ptarget))
{
*hashes_done = n - first_nonce + 1;
return true;
}
}
} while (n < max_nonce && !work_restart[thr_id].restart);
*hashes_done = n - first_nonce + 1;
pdata[19] = n;
return 0;
}
bool register_quark_algo( algo_gate_t* gate )
{
init_quark_ctx();
gate->optimizations = SSE2_OPT | AES_OPT;
gate->scanhash = (void*)&scanhash_quark;
gate->hash = (void*)&quarkhash;
gate->hash_alt = (void*)&quarkhash_alt;
return true;
};