Files
cpuminer-opt-gpu/algo/blake/blakecoin-4way.c
Jay D Dee 4e3f1b926f v23.14
2023-11-28 00:58:43 -05:00

206 lines
7.1 KiB
C

#include "blakecoin-gate.h"
#include "blake256-hash.h"
#include <string.h>
#include <stdint.h>
#include <memory.h>
#define rounds 8
#if defined (BLAKECOIN_16WAY)
int scanhash_blakecoin_16way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash32[8*16] __attribute__ ((aligned (64)));
uint32_t midstate_vars[16*16] __attribute__ ((aligned (64)));
__m512i block0_hash[8] __attribute__ ((aligned (64)));
__m512i block_buf[16] __attribute__ ((aligned (64)));
uint32_t lane_hash[8] __attribute__ ((aligned (32)));
uint32_t *hash32_d7 = (uint32_t*)&( ((__m512i*)hash32)[7] );
uint32_t *pdata = work->data;
const uint32_t *ptarget = work->target;
const uint32_t targ32_d7 = ptarget[7];
uint32_t phash[8] __attribute__ ((aligned (64))) =
{
0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
};
uint32_t n = pdata[19];
const uint32_t first_nonce = (const uint32_t) n;
const uint32_t last_nonce = max_nonce - 16;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
const __m512i sixteen = v512_32( 16 );
// Prehash first block
blake256_transform_le( phash, pdata, 512, 0, rounds );
block0_hash[0] = v512_32( phash[0] );
block0_hash[1] = v512_32( phash[1] );
block0_hash[2] = v512_32( phash[2] );
block0_hash[3] = v512_32( phash[3] );
block0_hash[4] = v512_32( phash[4] );
block0_hash[5] = v512_32( phash[5] );
block0_hash[6] = v512_32( phash[6] );
block0_hash[7] = v512_32( phash[7] );
// Build vectored second block, interleave last 16 bytes of data using
// unique nonces.
block_buf[0] = v512_32( pdata[16] );
block_buf[1] = v512_32( pdata[17] );
block_buf[2] = v512_32( pdata[18] );
block_buf[3] =
_mm512_set_epi32( n+15, n+14, n+13, n+12, n+11, n+10, n+ 9, n+ 8,
n+ 7, n+ 6, n+ 5, n+ 4, n+ 3, n+ 2, n +1, n );
// Partialy prehash second block without touching nonces in block_buf[3].
blake256_16way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
do {
blake256_16way_final_rounds_le( hash32, midstate_vars, block0_hash,
block_buf, rounds );
for ( int lane = 0; lane < 16; lane++ )
if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
{
extr_lane_16x32( lane_hash, hash32, lane, 256 );
if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
{
pdata[19] = n + lane;
submit_solution( work, lane_hash, mythr );
}
}
block_buf[3] = _mm512_add_epi32( block_buf[3], sixteen );
n += 16;
} while ( likely( (n < last_nonce) && !work_restart[thr_id].restart) );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
#elif defined (BLAKECOIN_8WAY)
int scanhash_blakecoin_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash32[8*8] __attribute__ ((aligned (64)));
uint32_t midstate_vars[16*8] __attribute__ ((aligned (32)));
__m256i block0_hash[8] __attribute__ ((aligned (32)));
__m256i block_buf[16] __attribute__ ((aligned (32)));
uint32_t lane_hash[8] __attribute__ ((aligned (32)));
uint32_t *hash32_d7 = (uint32_t*)&( ((__m256i*)hash32)[7] );
uint32_t *pdata = work->data;
const uint32_t *ptarget = work->target;
const uint32_t targ32_d7 = ptarget[7];
uint32_t phash[8] __attribute__ ((aligned (32))) =
{
0x6A09E667, 0xBB67AE85, 0x3C6EF372, 0xA54FF53A,
0x510E527F, 0x9B05688C, 0x1F83D9AB, 0x5BE0CD19
};
uint32_t n = pdata[19];
const uint32_t first_nonce = (const uint32_t) n;
const uint32_t last_nonce = max_nonce - 8;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
const __m256i eight = v256_32( 8 );
// Prehash first block
blake256_transform_le( phash, pdata, 512, 0, rounds );
block0_hash[0] = v256_32( phash[0] );
block0_hash[1] = v256_32( phash[1] );
block0_hash[2] = v256_32( phash[2] );
block0_hash[3] = v256_32( phash[3] );
block0_hash[4] = v256_32( phash[4] );
block0_hash[5] = v256_32( phash[5] );
block0_hash[6] = v256_32( phash[6] );
block0_hash[7] = v256_32( phash[7] );
// Build vectored second block, interleave last 16 bytes of data using
// unique nonces.
block_buf[0] = v256_32( pdata[16] );
block_buf[1] = v256_32( pdata[17] );
block_buf[2] = v256_32( pdata[18] );
block_buf[3] = _mm256_set_epi32( n+7, n+6, n+5, n+4, n+3, n+2, n+1, n );
// Partialy prehash second block without touching nonces in block_buf[3].
blake256_8way_round0_prehash_le( midstate_vars, block0_hash, block_buf );
do {
blake256_8way_final_rounds_le( hash32, midstate_vars, block0_hash,
block_buf, rounds );
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
{
extr_lane_8x32( lane_hash, hash32, lane, 256 );
if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
{
pdata[19] = n + lane;
submit_solution( work, lane_hash, mythr );
}
}
block_buf[3] = _mm256_add_epi32( block_buf[3], eight );
n += 8;
} while ( likely( (n < last_nonce) && !work_restart[thr_id].restart) );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
#elif defined (BLAKECOIN_4WAY)
blake256r8_4way_context blakecoin_4w_ctx;
void blakecoin_4way_hash(void *state, const void *input)
{
uint32_t vhash[8*4] __attribute__ ((aligned (64)));
blake256r8_4way_context ctx;
memcpy( &ctx, &blakecoin_4w_ctx, sizeof ctx );
blake256r8_4way_update( &ctx, input + (64<<2), 16 );
blake256r8_4way_close( &ctx, vhash );
dintrlv_4x32( state, state+32, state+64, state+96, vhash, 256 );
}
int scanhash_blakecoin_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t vdata[20*4] __attribute__ ((aligned (64)));
uint32_t hash[8*4] __attribute__ ((aligned (32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
uint32_t HTarget = ptarget[7];
uint32_t n = first_nonce;
__m128i *noncev = (__m128i*)vdata + 19; // aligned
int thr_id = mythr->id; // thr_id arg is deprecated
if ( opt_benchmark )
HTarget = 0x7f;
v128_bswap32_intrlv80_4x32( vdata, pdata );
blake256r8_4way_init( &blakecoin_4w_ctx );
blake256r8_4way_update( &blakecoin_4w_ctx, vdata, 64 );
do {
*noncev = v128_bswap32( _mm_set_epi32( n+3, n+2, n+1, n ) );
pdata[19] = n;
blakecoin_4way_hash( hash, vdata );
for ( int i = 0; i < 4; i++ )
if ( (hash+(i<<3))[7] <= HTarget && fulltest( hash+(i<<3), ptarget )
&& !opt_benchmark )
{
pdata[19] = n+i;
submit_solution( work, hash+(i<<3), mythr );
}
n += 4;
} while ( (n < max_nonce) && !work_restart[thr_id].restart );
*hashes_done = n - first_nonce + 1;
return 0;
}
#endif