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

v3.9.5

Browse Source
This commit is contained in:
Jay D Dee
2019-06-26 14:16:01 -04:00
parent d6e8d7a46e
commit 0d48d573ce
174 changed files with 1352 additions and 1556 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
algo/x13/drop.c
View File

@@ -175,7 +175,8 @@ static void droplp_hash_pok(void *output, uint32_t *pdata, const uint32_t versio
memcpy(output, hash, 32);
}
int scanhash_drop(int thr_id, struct work *work, uint32_t max_nonce, uint64_t *hashes_done)
int scanhash_drop( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t _ALIGN(64) hash[16];
uint32_t *pdata = work->data;
@@ -183,6 +184,7 @@ int scanhash_drop(int thr_id, struct work *work, uint32_t max_nonce, uint64_t *h
const uint32_t version = pdata[0] & (~POK_DATA_MASK);
const uint32_t first_nonce = pdata[19];
uint32_t nonce = first_nonce;
int thr_id = mythr->id; // thr_id arg is deprecated
#define tmpdata pdata
if (opt_benchmark)

37
algo/x13/phi1612-4way.c
View File

@@ -53,7 +53,7 @@ void phi1612_4way_hash( void *state, const void *input )
jh512_4way_close( &ctx.jh, vhash );
// Serial to the end
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// Cubehash
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
@@ -109,52 +109,41 @@ void phi1612_4way_hash( void *state, const void *input )
memcpy( state+96, hash3, 32 );
}
int scanhash_phi1612_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_phi1612_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
uint32_t _ALIGN(64) endiandata[20];
uint32_t n = first_nonce;
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
__m256i *noncev = (__m256i*)vdata + 9; // aligned
int thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
if ( opt_benchmark )
( (uint32_t*)ptarget )[7] = 0x0cff;
for ( int k = 0; k < 19; k++ )
be32enc( &endiandata[k], pdata[k] );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_bswap_intrlv80_4x64( vdata, pdata );
do {
be32enc( noncep, n );
be32enc( noncep+2, n+1 );
be32enc( noncep+4, n+2 );
be32enc( noncep+6, n+3 );
*noncev = mm256_intrlv_blend_32( mm256_bswap_32(
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ) ), *noncev );
phi1612_4way_hash( hash, vdata );
pdata[19] = n;
for ( int i = 0; i < 4; i++ )
if ( (hash+(i<<3))[7] <= Htarg && fulltest( hash+(i<<3), ptarget ) )
if ( (hash+(i<<3))[7] <= Htarg )
if ( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
{
pdata[19] = n+i;
nonces[ num_found++ ] = n+i;
work_set_target_ratio( work, hash+(i<<3) );
submit_lane_solution( work, hash+(i<<3), mythr, i );
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce )
&& !work_restart[thr_id].restart );
} while ( ( n < max_nonce ) && !work_restart[thr_id].restart );
*hashes_done = n - first_nonce + 1;
return num_found;
return 0;
}
#endif

8
algo/x13/phi1612-gate.h
View File

@@ -14,8 +14,8 @@ bool register_phi1612_algo( algo_gate_t* gate );
void phi1612_4way_hash( void *state, const void *input );
int scanhash_phi1612_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_phi1612_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_phi1612_4way_ctx();
@@ -23,8 +23,8 @@ void init_phi1612_4way_ctx();
void phi1612_hash( void *state, const void *input );
int scanhash_phi1612( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_phi1612( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_phi1612_ctx();

5
algo/x13/phi1612.c
View File

@@ -89,8 +89,8 @@ void phi1612_hash(void *output, const void *input)
memcpy(output, hash, 32);
}
int scanhash_phi1612( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_phi1612( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
@@ -98,6 +98,7 @@ int scanhash_phi1612( int thr_id, struct work *work, uint32_t max_nonce,
const uint32_t first_nonce = pdata[19];
uint32_t _ALIGN(64) endiandata[20];
uint32_t nonce = first_nonce;
int thr_id = mythr->id;
volatile uint8_t *restart = &(work_restart[thr_id].restart);
if (opt_benchmark)

21
algo/x13/skunk-4way.c
View File

@@ -33,7 +33,7 @@ void skunk_4way_hash( void *output, const void *input )
skein512_4way( &ctx.skein, input, 80 );
skein512_4way_close( &ctx.skein, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
cubehashUpdateDigest( &ctx.cube, (byte*) hash0, (const byte*)hash0, 64 );
memcpy( &ctx.cube, &skunk_4way_ctx.cube, sizeof(cubehashParam) );
@@ -73,8 +73,8 @@ void skunk_4way_hash( void *output, const void *input )
memcpy( output+96, hash3, 32 );
}
int scanhash_skunk_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_skunk_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
@@ -83,10 +83,9 @@ int scanhash_skunk_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *ptarget = work->target;
const uint32_t first_nonce = pdata[19];
uint32_t n = first_nonce;
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
const uint32_t Htarg = ptarget[7];
int thr_id = mythr->id; // thr_id arg is deprecated
volatile uint8_t *restart = &(work_restart[thr_id].restart);
if ( opt_benchmark )
@@ -95,7 +94,7 @@ int scanhash_skunk_4way( int thr_id, struct work *work, uint32_t max_nonce,
be32enc( &endiandata[k], pdata[k] );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
do
{
be32enc( noncep, n );
@@ -107,17 +106,17 @@ int scanhash_skunk_4way( int thr_id, struct work *work, uint32_t max_nonce,
pdata[19] = n;
for ( int i = 0; i < 4; i++ )
if ( (hash+(i<<3))[7] <= Htarg && fulltest( hash+(i<<3), ptarget ) )
if ( (hash+(i<<3))[7] <= Htarg )
if ( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
{
pdata[19] = n+i;
nonces[ num_found++ ] = n+i;
work_set_target_ratio( work, hash+(i<<3) );
submit_lane_solution( work, hash+(i<<3), mythr, i );
}
n +=4;
} while ( ( num_found == 0 ) && ( n < max_nonce ) && !(*restart) );
} while ( ( n < max_nonce ) && !(*restart) );
*hashes_done = n - first_nonce + 1;
return num_found;
return 0;
}
bool skunk_4way_thread_init()

8
algo/x13/skunk-gate.h
View File

@@ -14,8 +14,8 @@ bool register_skunk_algo( algo_gate_t* gate );
void skunk_4way_hash( void *state, const void *input );
int scanhash_skunk_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_skunk_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool skunk_4way_thread_init();
//void init_skunk_4way_ctx();
@@ -24,8 +24,8 @@ bool skunk_4way_thread_init();
void skunkhash( void *state, const void *input );
int scanhash_skunk( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_skunk( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool skunk_thread_init();

5
algo/x13/skunk.c
View File

@@ -38,8 +38,8 @@ void skunkhash( void *output, const void *input )
memcpy(output, hash, 32);
}
int scanhash_skunk( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_skunk( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t *pdata = work->data;
uint32_t *ptarget = work->target;
@@ -47,6 +47,7 @@ int scanhash_skunk( int thr_id, struct work *work, uint32_t max_nonce,
const uint32_t first_nonce = pdata[19];
uint32_t _ALIGN(64) endiandata[20];
uint32_t nonce = first_nonce;
int thr_id = mythr->id; // thr_id arg is deprecated
volatile uint8_t *restart = &(work_restart[thr_id].restart);
if ( opt_benchmark )

47
algo/x13/x13-4way.c
View File

@@ -74,7 +74,7 @@ void x13_4way_hash( void *state, const void *input )
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 3 Groestl
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -86,7 +86,7 @@ void x13_4way_hash( void *state, const void *input )
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// Parallel 4way 64 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// 4 Skein
skein512_4way( &ctx.skein, vhash, 64 );
@@ -101,16 +101,16 @@ void x13_4way_hash( void *state, const void *input )
keccak512_4way_close( &ctx.keccak, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 7 Luffa
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
mm256_intrlv_2x128( vhash, hash0, hash1, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, hash2, hash3, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhash, 512 );
mm256_intrlv_2x128( vhash, hash2, hash3, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhash, 512 );
// 8 Cubehash
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
@@ -138,13 +138,13 @@ void x13_4way_hash( void *state, const void *input )
sph_shavite512_close( &ctx.shavite, hash3 );
// 10 Simd
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
mm256_intrlv_2x128( vhash, hash0, hash1, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, hash2, hash3, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhash, 512 );
mm256_intrlv_2x128( vhash, hash2, hash3, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhash, 512 );
// 11 Echo
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -160,10 +160,10 @@ void x13_4way_hash( void *state, const void *input )
(const BitSequence *) hash3, 512 );
// 12 Hamsi parallel 4way 32 bit
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// 13 Fugue serial
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -184,8 +184,8 @@ void x13_4way_hash( void *state, const void *input )
memcpy( state+96, hash3, 32 );
}
int scanhash_x13_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_x13_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
@@ -194,9 +194,8 @@ int scanhash_x13_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
const uint32_t first_nonce = pdata[19];
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
int thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
@@ -207,7 +206,7 @@ int scanhash_x13_4way( int thr_id, struct work *work, uint32_t max_nonce,
swab32_array( endiandata, pdata, 20 );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
for ( int m=0; m < 6; m++ )
if ( Htarg <= htmax[m] )
@@ -224,21 +223,19 @@ int scanhash_x13_4way( int thr_id, struct work *work, uint32_t max_nonce,
pdata[19] = n;
for ( int i = 0; i < 4; i++ )
if ( ( ( (hash+(i<<3))[7] & mask ) == 0 )
&& fulltest( hash+(i<<3), ptarget ) )
if ( ( ( (hash+(i<<3))[7] & mask ) == 0 ) )
if ( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
{
pdata[19] = n+i;
nonces[ num_found++ ] = n+i;
work_set_target_ratio( work, hash+(i<<3) );
submit_lane_solution( work, hash+(i<<3), mythr, i );
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce )
&& !work_restart[thr_id].restart );
} while ( ( n < max_nonce ) && !work_restart[thr_id].restart );
break;
}
*hashes_done = n - first_nonce + 1;
return num_found;
return 0;
}
#endif

8
algo/x13/x13-gate.h
View File

@@ -14,8 +14,8 @@ bool register_x13_algo( algo_gate_t* gate );
void x13_4way_hash( void *state, const void *input );
int scanhash_x13_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_x13_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_x13_4way_ctx();
@@ -23,8 +23,8 @@ void init_x13_4way_ctx();
void x13hash( void *state, const void *input );
int scanhash_x13( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_x13( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_x13_ctx();

5
algo/x13/x13.c
View File

@@ -174,8 +174,8 @@ void x13hash(void *output, const void *input)
memcpy(output, hashB, 32);
}
int scanhash_x13(int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done)
int scanhash_x13( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(64)));
@@ -183,6 +183,7 @@ int scanhash_x13(int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
int thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = {

51
algo/x13/x13sm3-4way.c
View File

@@ -81,7 +81,7 @@ void x13sm3_4way_hash( void *state, const void *input )
bmw512_4way_close( &ctx.bmw, vhash );
// Serial
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// Groestl
update_and_final_groestl( &ctx.groestl, (char*)hash0, (char*)hash0, 512 );
@@ -93,7 +93,7 @@ void x13sm3_4way_hash( void *state, const void *input )
update_and_final_groestl( &ctx.groestl, (char*)hash3, (char*)hash3, 512 );
// Parallel 4way
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
// Skein
skein512_4way( &ctx.skein, vhash, 64 );
@@ -108,16 +108,16 @@ void x13sm3_4way_hash( void *state, const void *input )
keccak512_4way_close( &ctx.keccak, vhash );
// Serial to the end
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// Luffa
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
mm256_intrlv_2x128( vhash, hash0, hash1, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, hash2, hash3, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhash, 512 );
mm256_intrlv_2x128( vhash, hash2, hash3, 512 );
luffa_2way_init( &ctx.luffa, 512 );
luffa_2way_update_close( &ctx.luffa, vhash, vhash, 64 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhash, 512 );
// Cubehash
cubehashUpdateDigest( &ctx.cube, (byte*)hash0, (const byte*) hash0, 64 );
@@ -145,13 +145,13 @@ void x13sm3_4way_hash( void *state, const void *input )
sph_shavite512_close( &ctx.shavite, hash3 );
// Simd
mm256_interleave_2x128( vhash, hash0, hash1, 512 );
mm256_intrlv_2x128( vhash, hash0, hash1, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash0, hash1, vhash, 512 );
mm256_interleave_2x128( vhash, hash2, hash3, 512 );
mm256_dintrlv_2x128( hash0, hash1, vhash, 512 );
mm256_intrlv_2x128( vhash, hash2, hash3, 512 );
simd_2way_init( &ctx.simd, 512 );
simd_2way_update_close( &ctx.simd, vhash, vhash, 512 );
mm256_deinterleave_2x128( hash2, hash3, vhash, 512 );
mm256_dintrlv_2x128( hash2, hash3, vhash, 512 );
// Echo
update_final_echo( &ctx.echo, (BitSequence *)hash0,
@@ -166,7 +166,7 @@ void x13sm3_4way_hash( void *state, const void *input )
update_final_echo( &ctx.echo, (BitSequence *)hash3,
(const BitSequence *) hash3, 512 );
mm128_interleave_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
mm128_intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
// SM3 parallel 32 bit
uint32_t sm3_vhash[32*4] __attribute__ ((aligned (64)));
@@ -182,13 +182,13 @@ void x13sm3_4way_hash( void *state, const void *input )
sm3_4way( &ctx.sm3, vhash, 64 );
sm3_4way_close( &ctx.sm3, sm3_vhash );
mm128_deinterleave_4x32( hash0, hash1, hash2, hash3, sm3_vhash, 512 );
mm128_dintrlv_4x32( hash0, hash1, hash2, hash3, sm3_vhash, 512 );
// Hamsi parallel 4x32x2
mm256_interleave_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
mm256_intrlv_4x64( vhash, hash0, hash1, hash2, hash3, 512 );
hamsi512_4way( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
mm256_deinterleave_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
mm256_dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
// Fugue serial
sph_fugue512( &ctx.fugue, hash0, 64 );
@@ -209,8 +209,8 @@ void x13sm3_4way_hash( void *state, const void *input )
memcpy( state+96, hash3, 32 );
}
int scanhash_x13sm3_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done )
int scanhash_x13sm3_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
{
uint32_t hash[4*8] __attribute__ ((aligned (64)));
uint32_t vdata[24*4] __attribute__ ((aligned (64)));
@@ -219,9 +219,8 @@ int scanhash_x13sm3_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint32_t *ptarget = work->target;
uint32_t n = pdata[19];
const uint32_t first_nonce = pdata[19];
uint32_t *nonces = work->nonces;
int num_found = 0;
uint32_t *noncep = vdata + 73; // 9*8 + 1
int thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = { 0, 0xF, 0xFF,
0xFFF, 0xFFFF, 0x10000000 };
@@ -232,7 +231,7 @@ int scanhash_x13sm3_4way( int thr_id, struct work *work, uint32_t max_nonce,
swab32_array( endiandata, pdata, 20 );
uint64_t *edata = (uint64_t*)endiandata;
mm256_interleave_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
mm256_intrlv_4x64( (uint64_t*)vdata, edata, edata, edata, edata, 640 );
blake512_4way_init( &x13sm3_ctx_mid );
blake512_4way( &x13sm3_ctx_mid, vdata, 64 );
@@ -252,21 +251,19 @@ int scanhash_x13sm3_4way( int thr_id, struct work *work, uint32_t max_nonce,
pdata[19] = n;
for ( int i = 0; i < 4; i++ )
if ( ( ( (hash+(i<<3))[7] & mask ) == 0 )
&& fulltest( hash+(i<<3), ptarget ) )
if ( ( ( (hash+(i<<3))[7] & mask ) == 0 ) )
if ( fulltest( hash+(i<<3), ptarget ) && !opt_benchmark )
{
pdata[19] = n+i;
nonces[ num_found++ ] = n+i;
work_set_target_ratio( work, hash+(i<<3) );
submit_lane_solution( work, hash+(i<<3), mythr, i );
}
n += 4;
} while ( ( num_found == 0 ) && ( n < max_nonce )
&& !work_restart[thr_id].restart );
} while ( ( n < max_nonce ) && !work_restart[thr_id].restart );
break;
}
*hashes_done = n - first_nonce + 1;
return num_found;
return 0;
}
#endif

8
algo/x13/x13sm3-gate.h
View File

@@ -14,8 +14,8 @@ bool register_x13sm3_algo( algo_gate_t* gate );
void x13sm3_4way_hash( void *state, const void *input );
int scanhash_x13sm3_4way( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_x13sm3_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_x13sm3_4way_ctx();
@@ -23,8 +23,8 @@ void init_x13sm3_4way_ctx();
void x13sm3_hash( void *state, const void *input );
int scanhash_x13sm3( int thr_id, struct work *work, uint32_t max_nonce,
uint64_t *hashes_done );
int scanhash_x13sm3( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
void init_x13sm3_ctx();

5
algo/x13/x13sm3.c
View File

@@ -171,8 +171,8 @@ void x13sm3_hash(void *output, const void *input)
memcpy(output, hash, 32);
}
int scanhash_x13sm3( int thr_id, struct work *work,
uint32_t max_nonce, uint64_t *hashes_done)
int scanhash_x13sm3( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr)
{
uint32_t endiandata[20] __attribute__((aligned(64)));
uint32_t hash64[8] __attribute__((aligned(64)));
@@ -180,6 +180,7 @@ int scanhash_x13sm3( int thr_id, struct work *work,
uint32_t *ptarget = work->target;
uint32_t n = pdata[19] - 1;
const uint32_t first_nonce = pdata[19];
int thr_id = mythr->id; // thr_id arg is deprecated
const uint32_t Htarg = ptarget[7];
uint64_t htmax[] = {