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

v23.5

Browse Source
This commit is contained in:
Jay D Dee
2023-10-25 20:36:20 -04:00
parent 31c4dedf59
commit 160608cce5
180 changed files with 10318 additions and 13097 deletions
Download Patch File Download Diff File Expand all files Collapse all files

257
algo/sha/sha256dt.c
View File

@@ -4,28 +4,30 @@
#include <string.h>
#include <stdio.h>
#include "sha256-hash.h"
#include "sph_sha2.h"
#if defined(__AVX512F__) && defined(__AVX512VL__) && defined(__AVX512DQ__) && defined(__AVX512BW__)
#define SHA256DT_16WAY 1
#elif defined(__SHA__)
#define SHA256DT_SHA 1
#define SHA256DT_16X64 1
#elif defined(__x86_64__) && defined(__SHA__)
#define SHA256DT_X86_SHA256 1
#elif defined(__ARM_NEON) && defined(__ARM_FEATURE_SHA2)
#define SHA256DT_NEON_SHA2 1
#define SHA256DT_NEON_SHA256 1
#elif defined(__AVX2__)
#define SHA256DT_8WAY 1
#else
#define SHA256DT_4WAY 1
#define SHA256DT_8X64 1
#elif defined (__SSE2__) || defined(__ARM_NEON)
#define SHA256DT_4X64 1
#endif
// else ref, should never happen
static const uint32_t sha256dt_iv[8] __attribute__ ((aligned (32))) =
{
0xdfa9bf2c, 0xb72074d4, 0x6bb01122, 0xd338e869,
0xaa3ff126, 0x475bbf30, 0x8fd52e5b, 0x9f75c9ad
};
{
0xdfa9bf2c, 0xb72074d4, 0x6bb01122, 0xd338e869,
0xaa3ff126, 0x475bbf30, 0x8fd52e5b, 0x9f75c9ad
};
#if defined(SHA256DT_SHA)
#if defined(SHA256DT_X86_SHA256)
int scanhash_sha256dt_sha( struct work *work, uint32_t max_nonce,
int scanhash_sha256dt_x86_x2sha( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t block1a[16] __attribute__ ((aligned (64)));
@@ -44,8 +46,6 @@ int scanhash_sha256dt_sha( struct work *work, uint32_t max_nonce,
uint32_t n = first_nonce;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
const v128_t shuf_bswap32 =
v128_set64( 0x0c0d0e0f08090a0bULL, 0x0405060700010203ULL );
// hash first 64 byte block of data
sha256_transform_le( mstatea, pdata, sha256dt_iv );
@@ -106,13 +106,9 @@ int scanhash_sha256dt_sha( struct work *work, uint32_t max_nonce,
return 0;
}
#endif
#elif defined(SHA256DT_NEON_SHA256)
#if defined(SHA256DT_NEON_SHA2)
#pragma message "SHA256DT MEON SHA"
int scanhash_sha256dt_neon_sha2( struct work *work, uint32_t max_nonce,
int scanhash_sha256dt_neon_x2sha( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t block1a[16] __attribute__ ((aligned (64)));
@@ -121,8 +117,7 @@ int scanhash_sha256dt_neon_sha2( struct work *work, uint32_t max_nonce,
uint32_t block2b[16] __attribute__ ((aligned (64)));
uint32_t hasha[8] __attribute__ ((aligned (32)));
uint32_t hashb[8] __attribute__ ((aligned (32)));
uint32_t mstatea[8] __attribute__ ((aligned (32)));
uint32_t sstate[8] __attribute__ ((aligned (32)));
uint32_t mstate[8] __attribute__ ((aligned (32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
@@ -130,11 +125,9 @@ int scanhash_sha256dt_neon_sha2( struct work *work, uint32_t max_nonce,
uint32_t n = first_nonce;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
const v128_t shuf_bswap32 =
v128_set64( 0x0c0d0e0f08090a0bULL, 0x0405060700010203ULL );
// hash first 64 byte block of data
sha256_neon_transform_le( mstatea, pdata, sha256dt_iv );
sha256_neon_sha_transform_le( mstate, pdata, sha256dt_iv );
// fill & pad second bock without nonce
memcpy( block1a, pdata + 16, 12 );
@@ -156,10 +149,10 @@ int scanhash_sha256dt_neon_sha2( struct work *work, uint32_t max_nonce,
// Insert nonce for second block
block1a[3] = n;
block1b[3] = n+1;
sha256_neon2x_transform_le( block2a, block2b, block1a, block1b,
mstatea, mstatea );
sha256_neon_x2sha_transform_le( block2a, block2b, block1a, block1b,
mstate, mstate );
sha256_neon2x_transform_le( hasha, hashb, block2a, block2b,
sha256_neon_x2sha_transform_le( hasha, hashb, block2a, block2b,
sha256dt_iv, sha256dt_iv );
if ( unlikely( bswap_32( hasha[7] ) <= ptarget[7] ) )
@@ -190,11 +183,9 @@ int scanhash_sha256dt_neon_sha2( struct work *work, uint32_t max_nonce,
return 0;
}
#endif
#elif defined(SHA256DT_16X64)
#if defined(SHA256DT_16WAY)
int scanhash_sha256dt_16way( struct work *work, const uint32_t max_nonce,
int scanhash_sha256dt_16x64( struct work *work, const uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
__m512i block[16] __attribute__ ((aligned (128)));
@@ -241,7 +232,7 @@ int scanhash_sha256dt_16way( struct work *work, const uint32_t max_nonce,
buf[15] = v512_32( 0x480 ); // sha256dt funky bit count
// partially pre-expand & prehash second message block, avoiding the nonces
sha256_16way_prehash_3rounds( mstate2, mexp_pre, buf, mstate1 );
sha256_16x32_prehash_3rounds( mstate2, mexp_pre, buf, mstate1 );
// vectorize IV for second hash
istate[0] = v512_32( sha256dt_iv[0] );
@@ -260,8 +251,8 @@ int scanhash_sha256dt_16way( struct work *work, const uint32_t max_nonce,
do
{
sha256_16way_final_rounds( block, buf, mstate1, mstate2, mexp_pre );
if ( unlikely( sha256_16way_transform_le_short(
sha256_16x32_final_rounds( block, buf, mstate1, mstate2, mexp_pre );
if ( unlikely( sha256_16x32_transform_le_short(
hash32, block, istate, ptarget ) ) )
{
for ( int lane = 0; lane < 16; lane++ )
@@ -284,11 +275,9 @@ int scanhash_sha256dt_16way( struct work *work, const uint32_t max_nonce,
return 0;
}
#endif
#elif defined(SHA256DT_8X64)
#if defined(SHA256DT_8WAY)
int scanhash_sha256dt_8way( struct work *work, const uint32_t max_nonce,
int scanhash_sha256dt_8x64( struct work *work, const uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
__m256i vdata[32] __attribute__ ((aligned (64)));
@@ -335,15 +324,15 @@ int scanhash_sha256dt_8way( struct work *work, const uint32_t max_nonce,
istate[6] = v256_32( sha256dt_iv[6] );
istate[7] = v256_32( sha256dt_iv[7] );
sha256_8way_transform_le( mstate1, vdata, istate );
sha256_8x32_transform_le( mstate1, vdata, istate );
// Do 3 rounds on the first 12 bytes of the next block
sha256_8way_prehash_3rounds( mstate2, mexp_pre, vdata + 16, mstate1 );
sha256_8x32_prehash_3rounds( mstate2, mexp_pre, vdata + 16, mstate1 );
do
{
sha256_8way_final_rounds( block, vdata+16, mstate1, mstate2, mexp_pre );
if ( unlikely( sha256_8way_transform_le_short( hash32, block,
sha256_8x32_final_rounds( block, vdata+16, mstate1, mstate2, mexp_pre );
if ( unlikely( sha256_8x32_transform_le_short( hash32, block,
istate, ptarget ) ) )
{
for ( int lane = 0; lane < 8; lane++ )
@@ -366,19 +355,17 @@ int scanhash_sha256dt_8way( struct work *work, const uint32_t max_nonce,
return 0;
}
#endif
#elif defined(SHA256DT_4X64)
#if defined(SHA256DT_4WAY)
int scanhash_sha256dt_4way( struct work *work, const uint32_t max_nonce,
int scanhash_sha256dt_4x64( struct work *work, const uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
v128_t vdata[32] __attribute__ ((aligned (64)));
v128_t block[16] __attribute__ ((aligned (32)));
v128_t hash32[8] __attribute__ ((aligned (32)));
v128_t initstate[8] __attribute__ ((aligned (32)));
v128_t midstate[8] __attribute__ ((aligned (32)));
uint32_t lane_hash[8] __attribute__ ((aligned (32)));
v128_t iv[8] __attribute__ ((aligned (32)));
v128_t mhash[8] __attribute__ ((aligned (32)));
uint32_t lhash[8] __attribute__ ((aligned (32)));
uint32_t *hash32_d7 = (uint32_t*)&( hash32[7] );
uint32_t *pdata = work->data;
const uint32_t *ptarget = work->target;
@@ -406,33 +393,83 @@ int scanhash_sha256dt_4way( struct work *work, const uint32_t max_nonce,
block[15] = v128_32( 0x300 );
// initialize state
initstate[0] = v128_32( sha256dt_iv[0] );
initstate[1] = v128_32( sha256dt_iv[1] );
initstate[2] = v128_32( sha256dt_iv[2] );
initstate[3] = v128_32( sha256dt_iv[3] );
initstate[4] = v128_32( sha256dt_iv[4] );
initstate[5] = v128_32( sha256dt_iv[5] );
initstate[6] = v128_32( sha256dt_iv[6] );
initstate[7] = v128_32( sha256dt_iv[7] );
iv[0] = v128_32( sha256dt_iv[0] );
iv[1] = v128_32( sha256dt_iv[1] );
iv[2] = v128_32( sha256dt_iv[2] );
iv[3] = v128_32( sha256dt_iv[3] );
iv[4] = v128_32( sha256dt_iv[4] );
iv[5] = v128_32( sha256dt_iv[5] );
iv[6] = v128_32( sha256dt_iv[6] );
iv[7] = v128_32( sha256dt_iv[7] );
// hash first 64 bytes of data
sha256_4way_transform_le( midstate, vdata, initstate );
sha256_4x32_transform_le( mhash, vdata, iv );
/*
uint32_t m1 [8] __attribute__ ((aligned (32)));
uint32_t h1 [8] __attribute__ ((aligned (32)));
uint32_t b1 [16] __attribute__ ((aligned (32)));
uint32_t e16 [16] __attribute__ ((aligned (32)));
uint32_t *m4 = (uint32_t*)&midstate;
uint32_t *h4 = (uint32_t*)hash32;
sha256_transform_le( m1, pdata, sha256dt_iv );
memcpy( e16, pdata + 16, 12 );
e16[3] = n;
e16[4] = 0x80000000;
memset( &e16[5], 0, 40 );
e16[15] = 0x480; // funky bit count
b1[8] = 0x80000000;
memset( &b1[9], 0, 24 );
b1[9] = b1[10] = b1[11] = b1[12] = b1[13] = b1[14] = 0;
b1[15] = 0x300; // bit count
*/
do
{
sha256_4way_transform_le( block, vdata+16, midstate );
sha256_4way_transform_le( hash32, block, initstate );
sha256_4x32_transform_le( block, vdata+16, mhash );
v128_block_bswap32( hash32, hash32 );
//sha256_transform_le( b1, e16, m1 );
sha256_4x32_transform_le( hash32, block, iv );
/*
sha256_transform_le( h1, b1, sha256dt_iv );
printf("final hash1: %08x %08x %08x %08x %08x %08x %08x %08x\n",
h1[0],h1[1],h1[2],h1[3],h1[4],h1[5],h1[6],h1[7]);
printf("final hash4: %08x %08x %08x %08x %08x %08x %08x %08x\n",
h4[0],h4[4],h4[8],h4[12],h4[16],h4[20],h4[24],h4[28]);
casti_v128( h1,0 ) = v128_bswap32( casti_v128( h1,0 ) );
casti_v128( h1,1 ) = v128_bswap32( casti_v128( h1,1 ) );
*/
// v128_block_bswap32( hash32, hash32 );
/*
printf("bswap hash1: %08x %08x %08x %08x %08x %08x %08x %08x\n",
h1[0],h1[1],h1[2],h1[3],h1[4],h1[5],h1[6],h1[7]);
printf("bswap hash4: %08x %08x %08x %08x %08x %08x %08x %08x\n",
h4[0],h4[4],h4[8],h4[12],h4[16],h4[20],h4[24],h4[28]);
exit(0);
*/
for ( int lane = 0; lane < 4; lane++ )
if ( unlikely( hash32_d7[ lane ] <= targ32_d7 ) )
{
extr_lane_4x32( lane_hash, hash32, lane, 256 );
if ( likely( valid_hash( lane_hash, ptarget ) && !bench ) )
if ( unlikely( bswap_32( hash32_d7[ lane ] ) <= targ32_d7 ) )
{
pdata[19] = n + lane;
submit_solution( work, lane_hash, mythr );
extr_lane_4x32( lhash, hash32, lane, 256 );
casti_v128( lhash, 0 ) = v128_bswap32( casti_v128( lhash, 0 ) );
casti_v128( lhash, 1 ) = v128_bswap32( casti_v128( lhash, 1 ) );
if ( likely( valid_hash( lhash, ptarget ) && !bench ) )
{
pdata[19] = n + lane;
submit_solution( work, lhash, mythr );
}
}
}
*noncev = v128_add32( *noncev, four );
@@ -443,26 +480,84 @@ int scanhash_sha256dt_4way( struct work *work, const uint32_t max_nonce,
return 0;
}
#else
int scanhash_sha256dt_ref( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t block1a[16] __attribute__ ((aligned (32)));
uint32_t block2a[16] __attribute__ ((aligned (32)));
uint32_t hasha[8] __attribute__ ((aligned (32)));
uint32_t mstate[8] __attribute__ ((aligned (32)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
const uint32_t last_nonce = max_nonce - 2;
uint32_t n = first_nonce;
const int thr_id = mythr->id;
const bool bench = opt_benchmark;
// hash first 64 byte block of data
sha256_transform_le( mstate, pdata, sha256dt_iv );
// fill & pad second bock without nonce
memcpy( block1a, pdata + 16, 12 );
block1a[ 3] = 0;
block1a[ 4] = 0x80000000;
memset( block1a + 5, 0, 40 );
block1a[15] = 0x480; // funky bit count
// Pad third block
block2a[ 8] = 0x80000000;
memset( block2a + 9, 0, 24 );
block2a[15] = 0x300; // bit count
do
{
// Insert nonce for second block
block1a[3] = n;
sha256_transform_le( block2a, block1a, mstate );
sha256_transform_le( hasha, block2a, sha256dt_iv );
if ( unlikely( bswap_32( hasha[7] ) <= ptarget[7] ) )
{
casti_v128( hasha, 0 ) = v128_bswap32( casti_v128( hasha, 0 ) );
casti_v128( hasha, 1 ) = v128_bswap32( casti_v128( hasha, 1 ) );
if ( likely( valid_hash( hasha, ptarget ) && !bench ) )
{
pdata[19] = n;
submit_solution( work, hasha, mythr );
}
}
n += 1;
} while ( (n < last_nonce) && !work_restart[thr_id].restart );
pdata[19] = n;
*hashes_done = n - first_nonce;
return 0;
}
#endif
bool register_sha256dt_algo( algo_gate_t* gate )
{
gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT;
#if defined(SHA256DT_16WAY)
gate->scanhash = (void*)&scanhash_sha256dt_16way;
#elif defined(SHA256DT_SHA)
gate->optimizations = SSE2_OPT | AVX2_OPT | AVX512_OPT | NEON_OPT;
#if defined(SHA256DT_16X64)
gate->scanhash = (void*)&scanhash_sha256dt_16x64;
#elif defined(SHA256DT_X86_SHA256)
gate->optimizations = SHA_OPT;
gate->scanhash = (void*)&scanhash_sha256dt_sha;
#elif defined(SHA256DT_NEON_SHA2)
gate->scanhash = (void*)&scanhash_sha256dt_x86_x2sha;
#elif defined(SHA256DT_NEON_SHA256)
gate->optimizations = SHA_OPT;
gate->scanhash = (void*)&scanhash_sha256dt_neon_sha2;
#elif defined(SHA256DT_8WAY)
gate->scanhash = (void*)&scanhash_sha256dt_8way;
#elif defined(SHA256DT_4WAY)
gate->scanhash = (void*)&scanhash_sha256dt_4way;
gate->scanhash = (void*)&scanhash_sha256dt_neon_x2sha;
#elif defined(SHA256DT_8X64)
gate->scanhash = (void*)&scanhash_sha256dt_8x64;
#elif defined(SHA256DT_4X64)
gate->scanhash = (void*)&scanhash_sha256dt_4x64;
#else
gate->scanhash = (void*)&scanhash_sha256dt_ref;
#endif
return true;
}