popov/cpuminer-opt-gpu
1
0
Fork 0
mirror of https://github.com/JayDDee/cpuminer-opt.git synced 2025-09-17 23:44:27 +00:00
Code Issues Packages Projects Releases Wiki Activity

v3.21.3 Unreleased

Browse Source
This commit is contained in:
Jay D Dee
2023-03-13 03:20:13 -04:00
parent b339450898
commit c6bc9d67fb
49 changed files with 1126 additions and 1111 deletions
Download Patch File Download Diff File Expand all files Collapse all files

116
algo/x16/x21s-4way.c
View File

@@ -30,7 +30,7 @@ union _x21s_8way_context_overlay
typedef union _x21s_8way_context_overlay x21s_8way_context_overlay;
int x21s_8way_hash( void* output, const void* input, const int thrid )
int x21s_8way_hash( void* output, const void* input, int thrid )
{
uint32_t vhash[16*8] __attribute__ ((aligned (128)));
uint8_t shash[64*8] __attribute__ ((aligned (64)));
@@ -129,6 +129,66 @@ int x21s_8way_hash( void* output, const void* input, const int thrid )
return 1;
}
int scanhash_x21s_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr)
{
uint32_t hash[16*8] __attribute__ ((aligned (128)));
uint32_t vdata[20*8] __attribute__ ((aligned (64)));
uint32_t *hash7 = &hash[7<<3];
uint32_t lane_hash[8] __attribute__ ((aligned (64)));
uint32_t bedata1[2] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t Htarg = ptarget[7];
const uint32_t first_nonce = pdata[19];
uint32_t n = first_nonce;
const uint32_t last_nonce = max_nonce - 16;
const int thr_id = mythr->id;
__m512i *noncev = (__m512i*)vdata + 9; // aligned
volatile uint8_t *restart = &(work_restart[thr_id].restart);
const bool bench = opt_benchmark;
if ( bench ) ptarget[7] = 0x0cff;
bedata1[0] = bswap_32( pdata[1] );
bedata1[1] = bswap_32( pdata[2] );
static __thread uint32_t s_ntime = UINT32_MAX;
uint32_t ntime = bswap_32( pdata[17] );
if ( s_ntime != ntime )
{
x16_r_s_getAlgoString( (const uint8_t*)bedata1, x16r_hash_order );
s_ntime = ntime;
if ( opt_debug && !thr_id )
applog( LOG_INFO, "hash order %s (%08x)", x16r_hash_order, ntime );
}
x16r_8way_prehash( vdata, pdata );
*noncev = mm512_intrlv_blend_32( _mm512_set_epi32(
n+7, 0, n+6, 0, n+5, 0, n+4, 0,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
if ( x21s_8way_hash( hash, vdata, thr_id ) )
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( hash7[lane] <= Htarg ) )
{
extr_lane_8x32( lane_hash, hash, lane, 256 );
if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n + lane );
submit_solution( work, lane_hash, mythr );
}
}
*noncev = _mm512_add_epi32( *noncev,
m512_const1_64( 0x0000000800000000 ) );
n += 8;
} while ( likely( ( n < last_nonce ) && !(*restart) ) );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
bool x21s_8way_thread_init()
{
const int64_t ROW_LEN_INT64 = BLOCK_LEN_INT64 * 4; // nCols
@@ -155,7 +215,7 @@ union _x21s_4way_context_overlay
typedef union _x21s_4way_context_overlay x21s_4way_context_overlay;
int x21s_4way_hash( void* output, const void* input, const int thrid )
int x21s_4way_hash( void* output, const void* input, int thrid )
{
uint32_t vhash[16*4] __attribute__ ((aligned (64)));
uint8_t shash[64*4] __attribute__ ((aligned (64)));
@@ -231,6 +291,58 @@ int x21s_4way_hash( void* output, const void* input, const int thrid )
return 1;
}
int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr)
{
uint32_t hash[16*4] __attribute__ ((aligned (64)));
uint32_t vdata[20*4] __attribute__ ((aligned (64)));
uint32_t bedata1[2] __attribute__((aligned(64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
const uint32_t last_nonce = max_nonce - 4;
uint32_t n = first_nonce;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
__m256i *noncev = (__m256i*)vdata + 9; // aligned
volatile uint8_t *restart = &(work_restart[thr_id].restart);
if ( bench ) ptarget[7] = 0x0cff;
bedata1[0] = bswap_32( pdata[1] );
bedata1[1] = bswap_32( pdata[2] );
static __thread uint32_t s_ntime = UINT32_MAX;
uint32_t ntime = bswap_32( pdata[17] );
if ( s_ntime != ntime )
{
x16_r_s_getAlgoString( (const uint8_t*)bedata1, x16r_hash_order );
s_ntime = ntime;
if ( opt_debug && !thr_id )
applog( LOG_DEBUG, "hash order %s (%08x)", x16r_hash_order, ntime );
}
x16r_4way_prehash( vdata, pdata );
*noncev = mm256_intrlv_blend_32(
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
if ( x21s_4way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 4; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n+i );
submit_solution( work, hash+(i<<3), mythr );
}
*noncev = _mm256_add_epi32( *noncev,
m256_const1_64( 0x0000000400000000 ) );
n += 4;
} while ( likely( ( n < last_nonce ) && !(*restart) ) );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
bool x21s_4way_thread_init()
{
const int64_t ROW_LEN_INT64 = BLOCK_LEN_INT64 * 4; // nCols