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

v3.12.6.1

Browse Source
This commit is contained in:
Jay D Dee
2020-03-07 14:11:06 -05:00
parent c0aadbcc99
commit 6e8b8ed34f
29 changed files with 294 additions and 168 deletions
Download Patch File Download Diff File Expand all files Collapse all files

41
algo/scrypt/scrypt.c
View File

@@ -424,7 +424,7 @@ static bool scrypt_1024_1_1_256(const uint32_t *input, uint32_t *output,
}
#ifdef HAVE_SHA256_4WAY
static bool scrypt_1024_1_1_256_4way(const uint32_t *input,
static int scrypt_1024_1_1_256_4way(const uint32_t *input,
uint32_t *output, uint32_t *midstate, unsigned char *scratchpad, int N,
int thrid )
{
@@ -449,6 +449,8 @@ static bool scrypt_1024_1_1_256_4way(const uint32_t *input,
PBKDF2_SHA256_80_128_4way(tstate, ostate, W, W);
if ( work_restart[thrid].restart ) return 0;
for (i = 0; i < 32; i++)
for (k = 0; k < 4; k++)
X[k * 32 + i] = W[4 * i + k];
@@ -458,6 +460,8 @@ static bool scrypt_1024_1_1_256_4way(const uint32_t *input,
scrypt_core(X + 2 * 32, V, N);
scrypt_core(X + 3 * 32, V, N);
if ( work_restart[thrid].restart ) return 0;
for (i = 0; i < 32; i++)
for (k = 0; k < 4; k++)
W[4 * i + k] = X[k * 32 + i];
@@ -468,13 +472,13 @@ static bool scrypt_1024_1_1_256_4way(const uint32_t *input,
for (k = 0; k < 4; k++)
output[k * 8 + i] = W[4 * i + k];
return true;
return 1;
}
#endif /* HAVE_SHA256_4WAY */
#ifdef HAVE_SCRYPT_3WAY
static bool scrypt_1024_1_1_256_3way(const uint32_t *input,
static int scrypt_1024_1_1_256_3way(const uint32_t *input,
uint32_t *output, uint32_t *midstate, unsigned char *scratchpad, int N,
int thrid )
{
@@ -492,23 +496,23 @@ static bool scrypt_1024_1_1_256_3way(const uint32_t *input,
HMAC_SHA256_80_init(input + 20, tstate + 8, ostate + 8);
HMAC_SHA256_80_init(input + 40, tstate + 16, ostate + 16);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
PBKDF2_SHA256_80_128(tstate + 0, ostate + 0, input + 0, X + 0);
PBKDF2_SHA256_80_128(tstate + 8, ostate + 8, input + 20, X + 32);
PBKDF2_SHA256_80_128(tstate + 16, ostate + 16, input + 40, X + 64);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
scrypt_core_3way(X, V, N);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
PBKDF2_SHA256_128_32(tstate + 0, ostate + 0, X + 0, output + 0);
PBKDF2_SHA256_128_32(tstate + 8, ostate + 8, X + 32, output + 8);
PBKDF2_SHA256_128_32(tstate + 16, ostate + 16, X + 64, output + 16);
return true;
return 1;
}
#ifdef HAVE_SHA256_4WAY
@@ -539,13 +543,13 @@ static bool scrypt_1024_1_1_256_12way(const uint32_t *input,
HMAC_SHA256_80_init_4way(W + 128, tstate + 32, ostate + 32);
HMAC_SHA256_80_init_4way(W + 256, tstate + 64, ostate + 64);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
PBKDF2_SHA256_80_128_4way(tstate + 0, ostate + 0, W + 0, W + 0);
PBKDF2_SHA256_80_128_4way(tstate + 32, ostate + 32, W + 128, W + 128);
PBKDF2_SHA256_80_128_4way(tstate + 64, ostate + 64, W + 256, W + 256);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
for (j = 0; j < 3; j++)
for (i = 0; i < 32; i++)
@@ -557,7 +561,7 @@ static bool scrypt_1024_1_1_256_12way(const uint32_t *input,
scrypt_core_3way(X + 2 * 96, V, N);
scrypt_core_3way(X + 3 * 96, V, N);
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
for (j = 0; j < 3; j++)
for (i = 0; i < 32; i++)
@@ -573,14 +577,14 @@ static bool scrypt_1024_1_1_256_12way(const uint32_t *input,
for (k = 0; k < 4; k++)
output[32 * j + k * 8 + i] = W[128 * j + 4 * i + k];
return true;
return 1;
}
#endif /* HAVE_SHA256_4WAY */
#endif /* HAVE_SCRYPT_3WAY */
#ifdef HAVE_SCRYPT_6WAY
static bool scrypt_1024_1_1_256_24way( const uint32_t *input,
static int scrypt_1024_1_1_256_24way( const uint32_t *input,
uint32_t *output, uint32_t *midstate,
unsigned char *scratchpad, int N, int thrid )
{
@@ -607,13 +611,13 @@ static bool scrypt_1024_1_1_256_24way( const uint32_t *input,
HMAC_SHA256_80_init_8way( W + 256, tstate + 64, ostate + 64 );
HMAC_SHA256_80_init_8way( W + 512, tstate + 128, ostate + 128 );
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
PBKDF2_SHA256_80_128_8way( tstate + 0, ostate + 0, W + 0, W + 0 );
PBKDF2_SHA256_80_128_8way( tstate + 64, ostate + 64, W + 256, W + 256 );
PBKDF2_SHA256_80_128_8way( tstate + 128, ostate + 128, W + 512, W + 512 );
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
for ( j = 0; j < 3; j++ )
for ( i = 0; i < 32; i++ )
@@ -622,10 +626,13 @@ static bool scrypt_1024_1_1_256_24way( const uint32_t *input,
scrypt_core_6way( X + 0 * 32, V, N );
scrypt_core_6way( X + 6 * 32, V, N );
scrypt_core_6way( X + 12 * 32, V, N );
if ( work_restart[thrid].restart ) return 0;
scrypt_core_6way( X + 12 * 32, V, N );
scrypt_core_6way( X + 18 * 32, V, N );
if ( work_restart[thrid].restart ) return false;
if ( work_restart[thrid].restart ) return 0;
for ( j = 0; j < 3; j++ )
for ( i = 0; i < 32; i++ )
@@ -641,7 +648,7 @@ static bool scrypt_1024_1_1_256_24way( const uint32_t *input,
for ( k = 0; k < 8; k++ )
output[8 * 8 * j + k * 8 + i] = W[8 * 32 * j + 8 * i + k];
return true;
return 1;
}
#endif /* HAVE_SCRYPT_6WAY */

9
algo/x16/hex.c
View File

@@ -77,7 +77,7 @@ typedef union _hex_context_overlay hex_context_overlay;
static __thread x16r_context_overlay hex_ctx;
void hex_hash( void* output, const void* input )
int hex_hash( void* output, const void* input, int thrid )
{
uint32_t _ALIGN(128) hash[16];
x16r_context_overlay ctx;
@@ -214,11 +214,15 @@ void hex_hash( void* output, const void* input )
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
break;
}
if ( work_restart[thrid].restart ) return 0;
algo = (uint8_t)hash[0] % X16R_HASH_FUNC_COUNT;
in = (void*) hash;
size = 64;
}
memcpy(output, hash, 32);
return 1;
}
int scanhash_hex( struct work *work, uint32_t max_nonce,
@@ -286,8 +290,7 @@ int scanhash_hex( struct work *work, uint32_t max_nonce,
do
{
edata[19] = nonce;
hex_hash( hash32, edata );
if ( hex_hash( hash32, edata, thr_id ) );
if ( unlikely( valid_hash( hash32, ptarget ) && !bench ) )
{
be32enc( &pdata[19], nonce );

35
algo/x16/x16r-4way.c
View File

@@ -80,7 +80,7 @@ void x16r_8way_prehash( void *vdata, void *pdata )
// Called by wrapper hash function to optionally continue hashing and
// convert to final hash.
void x16r_8way_hash_generic( void* output, const void* input )
int x16r_8way_hash_generic( void* output, const void* input, int thrid )
{
uint32_t vhash[20*8] __attribute__ ((aligned (128)));
uint32_t hash0[20] __attribute__ ((aligned (64)));
@@ -424,6 +424,9 @@ void x16r_8way_hash_generic( void* output, const void* input )
hash7, vhash );
break;
}
if ( work_restart[thrid].restart ) return 0;
size = 64;
}
@@ -435,14 +438,17 @@ void x16r_8way_hash_generic( void* output, const void* input )
memcpy( output+320, hash5, 64 );
memcpy( output+384, hash6, 64 );
memcpy( output+448, hash7, 64 );
return 1;
}
// x16-r,-s,-rt wrapper called directly by scanhash to repackage 512 bit
// hash to 256 bit final hash.
void x16r_8way_hash( void* output, const void* input )
int x16r_8way_hash( void* output, const void* input, int thrid )
{
uint8_t hash[64*8] __attribute__ ((aligned (128)));
x16r_8way_hash_generic( hash, input );
if ( !x16r_8way_hash_generic( hash, input, thrid ) )
return 0;
memcpy( output, hash, 32 );
memcpy( output+32, hash+64, 32 );
@@ -452,7 +458,9 @@ void x16r_8way_hash( void* output, const void* input )
memcpy( output+160, hash+320, 32 );
memcpy( output+192, hash+384, 32 );
memcpy( output+224, hash+448, 32 );
}
return 1;
}
// x16r only
int scanhash_x16r_8way( struct work *work, uint32_t max_nonce,
@@ -492,8 +500,7 @@ int scanhash_x16r_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x16r_8way_hash( hash, vdata );
if( x16r_8way_hash( hash, vdata, thr_id ) );
for ( int i = 0; i < 8; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{
@@ -565,7 +572,7 @@ void x16r_4way_prehash( void *vdata, void *pdata )
}
}
void x16r_4way_hash_generic( void* output, const void* input )
int x16r_4way_hash_generic( void* output, const void* input, int thrid )
{
uint32_t vhash[20*4] __attribute__ ((aligned (128)));
uint32_t hash0[20] __attribute__ ((aligned (64)));
@@ -794,23 +801,31 @@ void x16r_4way_hash_generic( void* output, const void* input )
dintrlv_4x64_512( hash0, hash1, hash2, hash3, vhash );
break;
}
if ( work_restart[thrid].restart ) return 0;
size = 64;
}
memcpy( output, hash0, 64 );
memcpy( output+64, hash1, 64 );
memcpy( output+128, hash2, 64 );
memcpy( output+192, hash3, 64 );
return 1;
}
void x16r_4way_hash( void* output, const void* input )
int x16r_4way_hash( void* output, const void* input, int thrid )
{
uint8_t hash[64*4] __attribute__ ((aligned (64)));
x16r_4way_hash_generic( hash, input );
if ( !x16r_4way_hash_generic( hash, input, thrid ) )
return 0;
memcpy( output, hash, 32 );
memcpy( output+32, hash+64, 32 );
memcpy( output+64, hash+128, 32 );
memcpy( output+96, hash+192, 32 );
return 1;
}
int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
@@ -849,7 +864,7 @@ int scanhash_x16r_4way( struct work *work, uint32_t max_nonce,
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x16r_4way_hash( hash, vdata );
if ( x16r_4way_hash( hash, vdata, thr_id ) );
for ( int i = 0; i < 4; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{

24
algo/x16/x16r-gate.h
View File

@@ -131,8 +131,8 @@ typedef union _x16r_8way_context_overlay x16r_8way_context_overlay;
extern __thread x16r_8way_context_overlay x16r_ctx;
void x16r_8way_prehash( void *, void * );
void x16r_8way_hash_generic( void *, const void * );
void x16r_8way_hash( void *, const void * );
int x16r_8way_hash_generic( void *, const void *, int );
int x16r_8way_hash( void *, const void *, int );
int scanhash_x16r_8way( struct work *, uint32_t ,
uint64_t *, struct thr_info * );
extern __thread x16r_8way_context_overlay x16r_ctx;
@@ -166,8 +166,8 @@ typedef union _x16r_4way_context_overlay x16r_4way_context_overlay;
extern __thread x16r_4way_context_overlay x16r_ctx;
void x16r_4way_prehash( void *, void * );
void x16r_4way_hash_generic( void *, const void * );
void x16r_4way_hash( void *, const void * );
int x16r_4way_hash_generic( void *, const void *, int );
int x16r_4way_hash( void *, const void *, int );
int scanhash_x16r_4way( struct work *, uint32_t,
uint64_t *, struct thr_info * );
extern __thread x16r_4way_context_overlay x16r_ctx;
@@ -205,26 +205,26 @@ typedef union _x16r_context_overlay x16r_context_overlay;
extern __thread x16r_context_overlay x16_ctx;
void x16r_prehash( void *, void * );
void x16r_hash_generic( void *, const void * );
void x16r_hash( void *, const void * );
int x16r_hash_generic( void *, const void *, int );
int x16r_hash( void *, const void *, int );
int scanhash_x16r( struct work *, uint32_t, uint64_t *, struct thr_info * );
// x16Rv2
#if defined(X16RV2_8WAY)
void x16rv2_8way_hash( void *state, const void *input );
int x16rv2_8way_hash( void *state, const void *input, int thrid );
int scanhash_x16rv2_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#elif defined(X16RV2_4WAY)
void x16rv2_4way_hash( void *state, const void *input );
int x16rv2_4way_hash( void *state, const void *input, int thrid );
int scanhash_x16rv2_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#else
void x16rv2_hash( void *state, const void *input );
int x16rv2_hash( void *state, const void *input, int thr_id );
int scanhash_x16rv2( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
@@ -254,21 +254,21 @@ int scanhash_x16rt( struct work *work, uint32_t max_nonce,
// x21s
#if defined(X16R_8WAY)
void x21s_8way_hash( void *state, const void *input );
int x21s_8way_hash( void *state, const void *input, int thrid );
int scanhash_x21s_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool x21s_8way_thread_init();
#elif defined(X16R_4WAY)
void x21s_4way_hash( void *state, const void *input );
int x21s_4way_hash( void *state, const void *input, int thrid );
int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool x21s_4way_thread_init();
#else
void x21s_hash( void *state, const void *input );
int x21s_hash( void *state, const void *input, int thr_id );
int scanhash_x21s( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
bool x21s_thread_init();

17
algo/x16/x16r.c
View File

@@ -48,7 +48,7 @@ void x16r_prehash( void *edata, void *pdata )
}
}
void x16r_hash_generic( void* output, const void* input )
int x16r_hash_generic( void* output, const void* input, int thrid )
{
uint32_t _ALIGN(128) hash[16];
x16r_context_overlay ctx;
@@ -178,18 +178,24 @@ void x16r_hash_generic( void* output, const void* input )
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
break;
}
if ( work_restart[thrid].restart ) return 0;
in = (void*) hash;
size = 64;
}
memcpy( output, hash, 64 );
return true;
}
void x16r_hash( void* output, const void* input )
int x16r_hash( void* output, const void* input, int thrid )
{
uint8_t hash[64] __attribute__ ((aligned (64)));
x16r_hash_generic( hash, input );
if ( !x16r_hash_generic( hash, input, thrid ) )
return 0;
memcpy( output, hash, 32 );
memcpy( output, hash, 32 );
return 1;
}
int scanhash_x16r( struct work *work, uint32_t max_nonce,
@@ -223,8 +229,7 @@ int scanhash_x16r( struct work *work, uint32_t max_nonce,
do
{
edata[19] = nonce;
x16r_hash( hash32, edata );
if ( x16r_hash( hash32, edata, thr_id ) )
if ( unlikely( valid_hash( hash32, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( nonce );

5
algo/x16/x16rt-4way.c
View File

@@ -41,8 +41,7 @@ int scanhash_x16rt_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x16r_8way_hash( hash, vdata );
if ( x16r_8way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 8; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{
@@ -95,7 +94,7 @@ int scanhash_x16rt_4way( struct work *work, uint32_t max_nonce,
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x16r_4way_hash( hash, vdata );
if ( x16r_4way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 4; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{

3
algo/x16/x16rt.c
View File

@@ -36,8 +36,7 @@ int scanhash_x16rt( struct work *work, uint32_t max_nonce,
do
{
edata[19] = nonce;
x16r_hash( hash32, edata );
if ( x16r_hash( hash32, edata, thr_id ) )
if ( valid_hash( hash32, ptarget ) && !bench )
{
pdata[19] = bswap_32( nonce );

17
algo/x16/x16rv2-4way.c
View File

@@ -65,7 +65,7 @@ union _x16rv2_8way_context_overlay
typedef union _x16rv2_8way_context_overlay x16rv2_8way_context_overlay;
static __thread x16rv2_8way_context_overlay x16rv2_ctx;
void x16rv2_8way_hash( void* output, const void* input )
int x16rv2_8way_hash( void* output, const void* input, int thrid )
{
uint32_t vhash[24*8] __attribute__ ((aligned (128)));
uint32_t hash0[24] __attribute__ ((aligned (64)));
@@ -563,6 +563,9 @@ void x16rv2_8way_hash( void* output, const void* input )
hash7, vhash );
break;
}
if ( work_restart[thrid].restart ) return 0;
size = 64;
}
@@ -574,6 +577,7 @@ void x16rv2_8way_hash( void* output, const void* input )
memcpy( output+160, hash5, 32 );
memcpy( output+192, hash6, 32 );
memcpy( output+224, hash7, 32 );
return 1;
}
int scanhash_x16rv2_8way( struct work *work, uint32_t max_nonce,
@@ -669,8 +673,7 @@ int scanhash_x16rv2_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x16rv2_8way_hash( hash, vdata );
if ( x16rv2_8way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 8; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{
@@ -718,7 +721,7 @@ inline void padtiger512( uint32_t* hash )
for ( int i = 6; i < 16; i++ ) hash[i] = 0;
}
void x16rv2_4way_hash( void* output, const void* input )
int x16rv2_4way_hash( void* output, const void* input, int thrid )
{
uint32_t hash0[20] __attribute__ ((aligned (64)));
uint32_t hash1[20] __attribute__ ((aligned (64)));
@@ -1023,12 +1026,16 @@ void x16rv2_4way_hash( void* output, const void* input )
dintrlv_4x64( hash0, hash1, hash2, hash3, vhash, 512 );
break;
}
if ( work_restart[thrid].restart ) return 0;
size = 64;
}
memcpy( output, hash0, 32 );
memcpy( output+32, hash1, 32 );
memcpy( output+64, hash2, 32 );
memcpy( output+96, hash3, 32 );
return 1;
}
int scanhash_x16rv2_4way( struct work *work, uint32_t max_nonce,
@@ -1119,7 +1126,7 @@ int scanhash_x16rv2_4way( struct work *work, uint32_t max_nonce,
do
{
x16rv2_4way_hash( hash, vdata );
if ( x16rv2_4way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 4; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{

9
algo/x16/x16rv2.c
View File

@@ -67,7 +67,7 @@ inline void padtiger512(uint32_t* hash) {
for (int i = (24/4); i < (64/4); i++) hash[i] = 0;
}
void x16rv2_hash( void* output, const void* input )
int x16rv2_hash( void* output, const void* input, int thrid )
{
uint32_t _ALIGN(128) hash[16];
x16rv2_context_overlay ctx;
@@ -180,10 +180,14 @@ void x16rv2_hash( void* output, const void* input )
SHA512_Final( (unsigned char*) hash, &ctx.sha512 );
break;
}
if ( work_restart[thrid].restart ) return 0;
in = (void*) hash;
size = 64;
}
memcpy(output, hash, 32);
return 1;
}
int scanhash_x16rv2( struct work *work, uint32_t max_nonce,
@@ -221,8 +225,7 @@ int scanhash_x16rv2( struct work *work, uint32_t max_nonce,
do
{
edata[19] = nonce;
x16rv2_hash( hash32, edata );
if ( x16rv2_hash( hash32, edata, thr_id ) )
if ( unlikely( valid_hash( hash32, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( nonce );

21
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;
void x21s_8way_hash( void* output, const void* input )
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)));
@@ -44,7 +44,8 @@ void x21s_8way_hash( void* output, const void* input )
uint32_t *hash7 = (uint32_t*)( shash+448 );
x21s_8way_context_overlay ctx;
x16r_8way_hash_generic( shash, input );
if ( !x16r_8way_hash_generic( shash, input, thrid ) )
return 0;
intrlv_8x32_512( vhash, hash0, hash1, hash2, hash3, hash4, hash5, hash6,
hash7 );
@@ -124,6 +125,8 @@ void x21s_8way_hash( void* output, const void* input )
sha256_8way_init( &ctx.sha256 );
sha256_8way_update( &ctx.sha256, vhash, 64 );
sha256_8way_close( &ctx.sha256, output );
return 1;
}
int scanhash_x21s_8way( struct work *work, uint32_t max_nonce,
@@ -166,8 +169,7 @@ int scanhash_x21s_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x21s_8way_hash( hash, vdata );
if ( x21s_8way_hash( hash, vdata, thr_id ) )
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( hash7[lane] <= Htarg ) )
{
@@ -215,7 +217,7 @@ union _x21s_4way_context_overlay
typedef union _x21s_4way_context_overlay x21s_4way_context_overlay;
void x21s_4way_hash( void* output, const void* input )
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)));
@@ -225,8 +227,9 @@ void x21s_4way_hash( void* output, const void* input )
uint32_t *hash2 = (uint32_t*)( shash+128 );
uint32_t *hash3 = (uint32_t*)( shash+192 );
x16r_4way_hash_generic( shash, input );
if ( !x16r_4way_hash_generic( shash, input, thrid ) )
return 0;
intrlv_4x32( vhash, hash0, hash1, hash2, hash3, 512 );
haval256_5_4way_init( &ctx.haval );
@@ -299,6 +302,8 @@ void x21s_4way_hash( void* output, const void* input )
dintrlv_4x32( output, output+32, output+64,output+96, vhash, 256 );
#endif
return 1;
}
int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
@@ -337,7 +342,7 @@ int scanhash_x21s_4way( struct work *work, uint32_t max_nonce,
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x21s_4way_hash( hash, vdata );
if ( x21s_4way_hash( hash, vdata, thr_id ) )
for ( int i = 0; i < 4; i++ )
if ( unlikely( valid_hash( hash + (i<<3), ptarget ) && !bench ) )
{

10
algo/x16/x21s.c
View File

@@ -27,12 +27,13 @@ union _x21s_context_overlay
};
typedef union _x21s_context_overlay x21s_context_overlay;
void x21s_hash( void* output, const void* input )
int x21s_hash( void* output, const void* input, int thrid )
{
uint32_t _ALIGN(128) hash[16];
x21s_context_overlay ctx;
x16r_hash_generic( hash, input );
if ( !x16r_hash_generic( hash, input, thrid ) )
return 0;
sph_haval256_5_init( &ctx.haval );
sph_haval256_5( &ctx.haval, (const void*) hash, 64) ;
@@ -54,6 +55,8 @@ void x21s_hash( void* output, const void* input )
SHA256_Final( (unsigned char*)hash, &ctx.sha256 );
memcpy( output, hash, 32 );
return 1;
}
int scanhash_x21s( struct work *work, uint32_t max_nonce,
@@ -87,8 +90,7 @@ int scanhash_x21s( struct work *work, uint32_t max_nonce,
do
{
edata[19] = nonce;
x21s_hash( hash32, edata );
if ( x21s_hash( hash32, edata, thr_id ) )
if ( unlikely( valid_hash( hash32, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( nonce );

33
algo/x22/x22i-4way.c
View File

@@ -62,7 +62,7 @@ union _x22i_8way_ctx_overlay
};
typedef union _x22i_8way_ctx_overlay x22i_8way_ctx_overlay;
void x22i_8way_hash( void *output, const void *input )
int x22i_8way_hash( void *output, const void *input, int thrid )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
uint64_t vhashA[8*8] __attribute__ ((aligned (64)));
@@ -129,6 +129,8 @@ void x22i_8way_hash( void *output, const void *input )
keccak512_8way_update( &ctx.keccak, vhash, 64 );
keccak512_8way_close( &ctx.keccak, vhash );
if ( work_restart[thrid].restart ) return 0;
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
luffa512_4way_full( &ctx.luffa, vhashA, vhashA, 64 );
@@ -214,6 +216,8 @@ void x22i_8way_hash( void *output, const void *input )
#endif
if ( work_restart[thrid].restart ) return 0;
hamsi512_8way_init( &ctx.hamsi );
hamsi512_8way_update( &ctx.hamsi, vhash, 64 );
hamsi512_8way_close( &ctx.hamsi, vhash );
@@ -346,6 +350,8 @@ void x22i_8way_hash( void *output, const void *input )
sph_tiger (&ctx.tiger, (const void*) hash7, 64);
sph_tiger_close(&ctx.tiger, (void*) hashA7);
if ( work_restart[thrid].restart ) return 0;
memset( hash0, 0, 64 );
memset( hash1, 0, 64 );
memset( hash2, 0, 64 );
@@ -399,6 +405,8 @@ void x22i_8way_hash( void *output, const void *input )
sha256_8way_init( &ctx.sha256 );
sha256_8way_update( &ctx.sha256, vhash, 64 );
sha256_8way_close( &ctx.sha256, output );
return 1;
}
int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
@@ -428,8 +436,7 @@ int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x22i_8way_hash( hash, vdata );
if ( x22i_8way_hash( hash, vdata, thr_id ) )
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( ( hashd7[ lane ] <= targ32 ) && !bench ) )
{
@@ -437,7 +444,7 @@ int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
if ( likely( valid_hash( lane_hash, ptarget ) ) )
{
pdata[19] = bswap_32( n + lane );
submit_lane_solution( work, lane_hash, mythr, lane );
submit_solution( work, lane_hash, mythr );
}
}
*noncev = _mm512_add_epi32( *noncev,
@@ -524,7 +531,7 @@ union _x22i_4way_ctx_overlay
};
typedef union _x22i_4way_ctx_overlay x22i_ctx_overlay;
void x22i_4way_hash( void *output, const void *input )
int x22i_4way_hash( void *output, const void *input, int thrid )
{
uint64_t hash0[8*4] __attribute__ ((aligned (64)));
uint64_t hash1[8*4] __attribute__ ((aligned (64)));
@@ -563,6 +570,8 @@ void x22i_4way_hash( void *output, const void *input )
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
if ( work_restart[thrid].restart ) return false;
rintrlv_4x64_2x128( vhashA, vhashB, vhash, 512 );
luffa512_2way_full( &ctx.luffa, vhashA, vhashA, 64 );
@@ -591,6 +600,8 @@ void x22i_4way_hash( void *output, const void *input )
intrlv_4x64_512( vhash, hash0, hash1, hash2, hash3 );
if ( work_restart[thrid].restart ) return false;
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -636,6 +647,8 @@ void x22i_4way_hash( void *output, const void *input )
sha512_4way_close( &ctx.sha512, vhash );
dintrlv_4x64_512( &hash0[24], &hash1[24], &hash2[24], &hash3[24], vhash );
if ( work_restart[thrid].restart ) return false;
ComputeSingleSWIFFTX((unsigned char*)hash0, (unsigned char*)hashA0);
ComputeSingleSWIFFTX((unsigned char*)hash1, (unsigned char*)hashA1);
ComputeSingleSWIFFTX((unsigned char*)hash2, (unsigned char*)hashA2);
@@ -668,6 +681,8 @@ void x22i_4way_hash( void *output, const void *input )
sph_tiger (&ctx.tiger, (const void*) hash3, 64);
sph_tiger_close(&ctx.tiger, (void*) hashA3);
if ( work_restart[thrid].restart ) return false;
memset( hash0, 0, 64 );
memset( hash1, 0, 64 );
memset( hash2, 0, 64 );
@@ -700,8 +715,9 @@ void x22i_4way_hash( void *output, const void *input )
sha256_4way_init( &ctx.sha256 );
sha256_4way_update( &ctx.sha256, vhash, 64 );
sha256_4way_close( &ctx.sha256, output );
}
return 1;
}
int scanhash_x22i_4way( struct work* work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr )
@@ -729,8 +745,7 @@ int scanhash_x22i_4way( struct work* work, uint32_t max_nonce,
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x22i_4way_hash( hash, vdata );
if ( x22i_4way_hash( hash, vdata, thr_id ) )
for ( int lane = 0; lane < 4; lane++ )
if ( unlikely( hashd7[ lane ] <= targ32 && !bench ) )
{
@@ -738,7 +753,7 @@ int scanhash_x22i_4way( struct work* work, uint32_t max_nonce,
if ( valid_hash( lane_hash, ptarget ) )
{
pdata[19] = bswap_32( n + lane );
submit_lane_solution( work, lane_hash, mythr, lane );
submit_solution( work, lane_hash, mythr );
}
}
*noncev = _mm256_add_epi32( *noncev,

12
algo/x22/x22i-gate.h
View File

@@ -16,19 +16,19 @@ bool register_x22i_algo( algo_gate_t* gate );
#if defined(X22I_8WAY)
void x22i_8way_hash( void *state, const void *input );
int x22i_8way_hash( void *state, const void *input, int thrid );
int scanhash_x22i_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#elif defined(X22I_4WAY)
void x22i_4way_hash( void *state, const void *input );
int x22i_4way_hash( void *state, const void *input, int thrid );
int scanhash_x22i_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#else
void x22i_hash( void *state, const void *input );
int x22i_hash( void *state, const void *input, int thrid );
int scanhash_x22i( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
@@ -44,19 +44,19 @@ bool register_x25i_algo( algo_gate_t* gate );
#if defined(X25X_8WAY)
void x25x_8way_hash( void *state, const void *input );
int x25x_8way_hash( void *state, const void *input, int thrid );
int scanhash_x25x_8way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#elif defined(X25X_4WAY)
void x25x_4way_hash( void *state, const void *input );
int x25x_4way_hash( void *state, const void *input, int thrid );
int scanhash_x25x_4way( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );
#else
void x25x_hash( void *state, const void *input );
int x25x_hash( void *state, const void *input, int thrif );
int scanhash_x25x( struct work *work, uint32_t max_nonce,
uint64_t *hashes_done, struct thr_info *mythr );

12
algo/x22/x22i.c
View File

@@ -59,7 +59,7 @@ union _x22i_context_overlay
};
typedef union _x22i_context_overlay x22i_context_overlay;
void x22i_hash( void *output, const void *input )
int x22i_hash( void *output, const void *input, int thrid )
{
unsigned char hash[64 * 4] __attribute__((aligned(64))) = {0};
unsigned char hash2[65] __attribute__((aligned(64))) = {0};
@@ -95,6 +95,8 @@ void x22i_hash( void *output, const void *input )
sph_keccak512(&ctx.keccak, (const void*) hash, 64);
sph_keccak512_close(&ctx.keccak, hash);
if ( work_restart[thrid].restart ) return 0;
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)hash,
(const BitSequence*)hash, 64 );
@@ -121,6 +123,8 @@ void x22i_hash( void *output, const void *input )
sph_echo512_close( &ctx.echo, hash );
#endif
if ( work_restart[thrid].restart ) return 0;
sph_hamsi512_init(&ctx.hamsi);
sph_hamsi512(&ctx.hamsi, (const void*) hash, 64);
sph_hamsi512_close(&ctx.hamsi, hash);
@@ -143,6 +147,8 @@ void x22i_hash( void *output, const void *input )
ComputeSingleSWIFFTX((unsigned char*)hash, (unsigned char*)hash2);
if ( work_restart[thrid].restart ) return 0;
memset(hash, 0, 64);
sph_haval256_5_init(&ctx.haval);
sph_haval256_5(&ctx.haval,(const void*) hash2, 64);
@@ -165,6 +171,8 @@ void x22i_hash( void *output, const void *input )
SHA256_Final( (unsigned char*) hash, &ctx.sha256 );
memcpy(output, hash, 32);
return 1;
}
int scanhash_x22i( struct work *work, uint32_t max_nonce,
@@ -188,7 +196,7 @@ int scanhash_x22i( struct work *work, uint32_t max_nonce,
do
{
edata[19] = n;
x22i_hash( hash64, edata );
if ( x22i_hash( hash64, edata, thr_id ) );
if ( unlikely( valid_hash( hash64, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n );

36
algo/x22/x25x-4way.c
View File

@@ -94,7 +94,7 @@ union _x25x_8way_ctx_overlay
};
typedef union _x25x_8way_ctx_overlay x25x_8way_ctx_overlay;
void x25x_8way_hash( void *output, const void *input )
int x25x_8way_hash( void *output, const void *input, int thrid )
{
uint64_t vhash[8*8] __attribute__ ((aligned (128)));
unsigned char hash0[25][64] __attribute__((aligned(64))) = {0};
@@ -179,13 +179,15 @@ void x25x_8way_hash( void *output, const void *input )
jh512_8way_close( &ctx.jh, vhash );
dintrlv_8x64_512( hash0[4], hash1[4], hash2[4], hash3[4],
hash4[4], hash5[4], hash6[4], hash7[4], vhash );
keccak512_8way_init( &ctx.keccak );
keccak512_8way_update( &ctx.keccak, vhash, 64 );
keccak512_8way_close( &ctx.keccak, vhash );
dintrlv_8x64_512( hash0[5], hash1[5], hash2[5], hash3[5],
hash4[5], hash5[5], hash6[5], hash7[5], vhash );
if ( work_restart[thrid].restart ) return 0;
rintrlv_8x64_4x128( vhashA, vhashB, vhash, 512 );
luffa_4way_init( &ctx.luffa, 512 );
@@ -261,6 +263,7 @@ void x25x_8way_hash( void *output, const void *input )
intrlv_8x64_512( vhash, hash0[10], hash1[10], hash2[10], hash3[10],
hash4[10], hash5[10], hash6[10], hash7[10] );
#else
init_echo( &ctx.echo, 512 );
@@ -292,6 +295,8 @@ void x25x_8way_hash( void *output, const void *input )
#endif
if ( work_restart[thrid].restart ) return 0;
hamsi512_8way_init( &ctx.hamsi );
hamsi512_8way_update( &ctx.hamsi, vhash, 64 );
hamsi512_8way_close( &ctx.hamsi, vhash );
@@ -407,6 +412,8 @@ void x25x_8way_hash( void *output, const void *input )
sph_tiger (&ctx.tiger, (const void*) hash7[17], 64);
sph_tiger_close(&ctx.tiger, (void*) hash7[18]);
if ( work_restart[thrid].restart ) return 0;
intrlv_2x256( vhash, hash0[18], hash1[18], 256 );
LYRA2X_2WAY( vhash, 32, vhash, 32, 1, 4, 4 );
dintrlv_2x256( hash0[19], hash1[19], vhash, 256 );
@@ -468,6 +475,8 @@ void x25x_8way_hash( void *output, const void *input )
laneHash(512, (const BitSequence*)hash6[22], 512, (BitSequence*)hash6[23]);
laneHash(512, (const BitSequence*)hash7[22], 512, (BitSequence*)hash7[23]);
if ( work_restart[thrid].restart ) return 0;
x25x_shuffle( hash0 );
x25x_shuffle( hash1 );
x25x_shuffle( hash2 );
@@ -528,6 +537,8 @@ void x25x_8way_hash( void *output, const void *input )
blake2s_8way_init( &ctx.blake2s, 32 );
blake2s_8way_full_blocks( &ctx.blake2s, output, vhashX, 64*24 );
return 1;
}
int scanhash_x25x_8way( struct work *work, uint32_t max_nonce,
@@ -557,7 +568,7 @@ int scanhash_x25x_8way( struct work *work, uint32_t max_nonce,
n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x25x_8way_hash( hash, vdata );
if ( x25x_8way_hash( hash, vdata, thr_id ) );
for ( int lane = 0; lane < 8; lane++ )
if ( unlikely( ( hashd7[ lane ] <= targ32 ) && !bench ) )
@@ -566,7 +577,7 @@ int scanhash_x25x_8way( struct work *work, uint32_t max_nonce,
if ( likely( valid_hash( lane_hash, ptarget ) ) )
{
pdata[19] = bswap_32( n + lane );
submit_lane_solution( work, lane_hash, mythr, lane );
submit_solution( work, lane_hash, mythr );
}
}
*noncev = _mm512_add_epi32( *noncev,
@@ -654,7 +665,7 @@ union _x25x_4way_ctx_overlay
};
typedef union _x25x_4way_ctx_overlay x25x_4way_ctx_overlay;
void x25x_4way_hash( void *output, const void *input )
int x25x_4way_hash( void *output, const void *input, int thrid )
{
uint64_t vhash[8*4] __attribute__ ((aligned (128)));
unsigned char hash0[25][64] __attribute__((aligned(64))) = {0};
@@ -686,6 +697,8 @@ void x25x_4way_hash( void *output, const void *input )
jh512_4way_close( &ctx.jh, vhash );
dintrlv_4x64_512( hash0[4], hash1[4], hash2[4], hash3[4], vhash );
if ( work_restart[thrid].restart ) return 0;
keccak512_4way_init( &ctx.keccak );
keccak512_4way_update( &ctx.keccak, vhash, 64 );
keccak512_4way_close( &ctx.keccak, vhash );
@@ -738,6 +751,8 @@ void x25x_4way_hash( void *output, const void *input )
intrlv_4x64_512( vhash, hash0[10], hash1[10], hash2[10], hash3[10] );
if ( work_restart[thrid].restart ) return 0;
hamsi512_4way_init( &ctx.hamsi );
hamsi512_4way_update( &ctx.hamsi, vhash, 64 );
hamsi512_4way_close( &ctx.hamsi, vhash );
@@ -819,6 +834,8 @@ void x25x_4way_hash( void *output, const void *input )
LYRA2RE( (void*)hash3[19], 32, (const void*)hash3[18], 32,
(const void*)hash3[18], 32, 1, 4, 4 );
if ( work_restart[thrid].restart ) return 0;
sph_gost512_init(&ctx.gost);
sph_gost512 (&ctx.gost, (const void*) hash0[19], 64);
sph_gost512_close(&ctx.gost, (void*) hash0[20]);
@@ -850,6 +867,8 @@ void x25x_4way_hash( void *output, const void *input )
laneHash(512, (const BitSequence*)hash2[22], 512, (BitSequence*)hash2[23]);
laneHash(512, (const BitSequence*)hash3[22], 512, (BitSequence*)hash3[23]);
if ( work_restart[thrid].restart ) return 0;
x25x_shuffle( hash0 );
x25x_shuffle( hash1 );
x25x_shuffle( hash2 );
@@ -882,6 +901,8 @@ void x25x_4way_hash( void *output, const void *input )
blake2s_4way_init( &ctx.blake2s, 32 );
blake2s_4way_full_blocks( &ctx.blake2s, output, vhashX, 64*24 );
return 1;
}
int scanhash_x25x_4way( struct work* work, uint32_t max_nonce,
@@ -910,8 +931,7 @@ int scanhash_x25x_4way( struct work* work, uint32_t max_nonce,
_mm256_set_epi32( n+3, 0, n+2, 0, n+1, 0, n, 0 ), *noncev );
do
{
x25x_4way_hash( hash, vdata );
if ( x25x_4way_hash( hash, vdata, thr_id ) )
for ( int lane = 0; lane < 4; lane++ )
if ( unlikely( hashd7[ lane ] <= targ32 && !bench ) )
{
@@ -919,7 +939,7 @@ int scanhash_x25x_4way( struct work* work, uint32_t max_nonce,
if ( valid_hash( lane_hash, ptarget ) )
{
pdata[19] = bswap_32( n + lane );
submit_lane_solution( work, lane_hash, mythr, lane );
submit_solution( work, lane_hash, mythr );
}
}
*noncev = _mm256_add_epi32( *noncev,

14
algo/x22/x25x.c
View File

@@ -64,7 +64,7 @@ union _x25x_context_overlay
};
typedef union _x25x_context_overlay x25x_context_overlay;
void x25x_hash( void *output, const void *input )
int x25x_hash( void *output, const void *input, int thrid )
{
unsigned char hash[25][64] __attribute__((aligned(64))) = {0};
x25x_context_overlay ctx;
@@ -99,6 +99,8 @@ void x25x_hash( void *output, const void *input )
sph_keccak512(&ctx.keccak, (const void*) &hash[4], 64);
sph_keccak512_close(&ctx.keccak, &hash[5]);
if ( work_restart[thrid].restart ) return 0;
init_luffa( &ctx.luffa, 512 );
update_and_final_luffa( &ctx.luffa, (BitSequence*)&hash[6],
(const BitSequence*)&hash[5], 64 );
@@ -125,7 +127,9 @@ void x25x_hash( void *output, const void *input )
sph_echo512_close( &ctx.echo, &hash[10] );
#endif
sph_hamsi512_init(&ctx.hamsi);
if ( work_restart[thrid].restart ) return 0;
sph_hamsi512_init(&ctx.hamsi);
sph_hamsi512(&ctx.hamsi, (const void*) &hash[10], 64);
sph_hamsi512_close(&ctx.hamsi, &hash[11]);
@@ -151,6 +155,8 @@ void x25x_hash( void *output, const void *input )
sph_haval256_5(&ctx.haval,(const void*) &hash[16], 64);
sph_haval256_5_close(&ctx.haval,&hash[17]);
if ( work_restart[thrid].restart ) return 0;
sph_tiger_init(&ctx.tiger);
sph_tiger (&ctx.tiger, (const void*) &hash[17], 64);
sph_tiger_close(&ctx.tiger, (void*) &hash[18]);
@@ -199,6 +205,8 @@ void x25x_hash( void *output, const void *input )
blake2s_simple( (uint8_t*)&hash[24], (const void*)(&hash[0]), 64 * 24 );
memcpy(output, &hash[24], 32);
return 1;
}
int scanhash_x25x( struct work *work, uint32_t max_nonce,
@@ -222,7 +230,7 @@ int scanhash_x25x( struct work *work, uint32_t max_nonce,
do
{
edata[19] = n;
x25x_hash( hash64, edata );
if ( x25x_hash( hash64, edata, thr_id ) );
if ( unlikely( valid_hash( hash64, ptarget ) && !bench ) )
{
pdata[19] = bswap_32( n );

2
algo/yespower/benchmark.c
View File

@@ -79,7 +79,7 @@ int main(int argc, const char * const *argv)
for (i = 0; i < sizeof(src); i++)
src.u8[i] = i * 3;
if (yespower_tls(src.u8, sizeof(src), &params, &dst)) {
if (!yespower_tls(src.u8, sizeof(src), &params, &dst)) {
puts("FAILED");
return 1;
}

2
algo/yespower/yescrypt-r8g.c
View File

@@ -53,7 +53,7 @@ int scanhash_yespower_r8g( struct work *work, uint32_t max_nonce,
do {
yespower_tls( (unsigned char *)endiandata, params.perslen,
&params, (yespower_binary_t*)hash );
&params, (yespower_binary_t*)hash, thr_id );
if unlikely( valid_hash( hash, ptarget ) && !opt_benchmark )
{

20
algo/yespower/yespower-blake2b.c
View File

@@ -194,11 +194,13 @@ static int free_region(yespower_region_t *region)
#define restrict
#endif
/*
#ifdef __GNUC__
#define unlikely(exp) __builtin_expect(exp, 0)
#else
#define unlikely(exp) (exp)
#endif
*/
#ifdef __SSE__
#define PREFETCH(x, hint) _mm_prefetch((const char *)(x), (hint));
@@ -1113,7 +1115,7 @@ static void smix(uint8_t *B, size_t r, uint32_t N,
int yespower_b2b(yespower_local_t *local,
const uint8_t *src, size_t srclen,
const yespower_params_t *params,
yespower_binary_t *dst)
yespower_binary_t *dst, int thrid )
{
uint32_t N = params->N;
uint32_t r = params->r;
@@ -1168,17 +1170,25 @@ int yespower_b2b(yespower_local_t *local,
srclen = 0;
}
if ( work_restart[thrid].restart ) return false;
pbkdf2_blake2b_yp(init_hash, sizeof(init_hash), src, srclen, 1, B, 128);
if ( work_restart[thrid].restart ) return false;
memcpy(init_hash, B, sizeof(init_hash));
smix_1_0(B, r, N, V, XY, &ctx);
if ( work_restart[thrid].restart ) return false;
hmac_blake2b_yp_hash((uint8_t *)dst, B + B_size - 64, 64, init_hash, sizeof(init_hash));
/* Success! */
return 0;
return 1;
fail:
memset(dst, 0xff, sizeof(*dst));
return -1;
return 0;
}
/**
@@ -1189,7 +1199,7 @@ fail:
* Return 0 on success; or -1 on error.
*/
int yespower_b2b_tls(const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst)
const yespower_params_t *params, yespower_binary_t *dst, int thrid )
{
static __thread int initialized = 0;
static __thread yespower_local_t local;
@@ -1199,7 +1209,7 @@ int yespower_b2b_tls(const uint8_t *src, size_t srclen,
initialized = 1;
}
return yespower_b2b(&local, src, srclen, params, dst);
return yespower_b2b(&local, src, srclen, params, dst, thrid);
}
/*
int yespower_init_local(yespower_local_t *local)

13
algo/yespower/yespower-gate.c
View File

@@ -34,9 +34,10 @@ static yespower_params_t yespower_params;
// YESPOWER
void yespower_hash( const char *input, char *output, uint32_t len )
int yespower_hash( const char *input, char *output, uint32_t len, int thrid )
{
yespower_tls( input, len, &yespower_params, (yespower_binary_t*)output );
return yespower_tls( input, len, &yespower_params,
(yespower_binary_t*)output, thrid );
}
int scanhash_yespower( struct work *work, uint32_t max_nonce,
@@ -55,7 +56,7 @@ int scanhash_yespower( struct work *work, uint32_t max_nonce,
be32enc( &endiandata[k], pdata[k] );
endiandata[19] = n;
do {
yespower_hash( (char*)endiandata, (char*)vhash, 80 );
if ( yespower_hash( (char*)endiandata, (char*)vhash, 80, thr_id ) )
if unlikely( valid_hash( vhash, ptarget ) && !opt_benchmark )
{
be32enc( pdata+19, n );
@@ -70,9 +71,9 @@ int scanhash_yespower( struct work *work, uint32_t max_nonce,
// YESPOWER-B2B
void yespower_b2b_hash( const char *input, char *output, uint32_t len )
int yespower_b2b_hash( const char *input, char *output, uint32_t len, int thrid )
{
yespower_b2b_tls( input, len, &yespower_params, (yespower_binary_t*)output );
return yespower_b2b_tls( input, len, &yespower_params, (yespower_binary_t*)output, thrid );
}
int scanhash_yespower_b2b( struct work *work, uint32_t max_nonce,
@@ -91,7 +92,7 @@ int scanhash_yespower_b2b( struct work *work, uint32_t max_nonce,
be32enc( &endiandata[k], pdata[k] );
endiandata[19] = n;
do {
yespower_b2b_hash( (char*) endiandata, (char*) vhash, 80 );
if (yespower_b2b_hash( (char*) endiandata, (char*) vhash, 80, thr_id ) )
if unlikely( valid_hash( vhash, ptarget ) && !opt_benchmark )
{
be32enc( pdata+19, n );

25
algo/yespower/yespower-opt.c
View File

@@ -107,11 +107,13 @@
#define restrict
#endif
/*
#ifdef __GNUC__
#define unlikely(exp) __builtin_expect(exp, 0)
#else
#define unlikely(exp) (exp)
#endif
*/
#ifdef __SSE__
#define PREFETCH(x, hint) _mm_prefetch((const char *)(x), (hint));
@@ -1023,7 +1025,7 @@ static void smix(uint8_t *B, size_t r, uint32_t N,
int yespower(yespower_local_t *local,
const uint8_t *src, size_t srclen,
const yespower_params_t *params,
yespower_binary_t *dst)
yespower_binary_t *dst, int thrid )
{
yespower_version_t version = params->version;
uint32_t N = params->N;
@@ -1077,15 +1079,24 @@ int yespower(yespower_local_t *local,
if (version == YESPOWER_0_5) {
PBKDF2_SHA256(sha256, sizeof(sha256), src, srclen, 1,
B, B_size);
memcpy(sha256, B, sizeof(sha256));
if ( work_restart[thrid].restart ) return false;
memcpy(sha256, B, sizeof(sha256));
smix(B, r, N, V, XY, &ctx);
PBKDF2_SHA256(sha256, sizeof(sha256), B, B_size, 1,
if ( work_restart[thrid].restart ) return false;
PBKDF2_SHA256(sha256, sizeof(sha256), B, B_size, 1,
(uint8_t *)dst, sizeof(*dst));
if (pers) {
HMAC_SHA256_Buf(dst, sizeof(*dst), pers, perslen,
sha256);
SHA256_Buf(sha256, sizeof(sha256), (uint8_t *)dst);
if ( work_restart[thrid].restart ) return false;
SHA256_Buf(sha256, sizeof(sha256), (uint8_t *)dst);
}
} else {
ctx.S2 = S + 2 * Swidth_to_Sbytes1(Swidth);
@@ -1106,7 +1117,7 @@ int yespower(yespower_local_t *local,
}
/* Success! */
return 0;
return 1;
}
/**
@@ -1117,7 +1128,7 @@ int yespower(yespower_local_t *local,
* Return 0 on success; or -1 on error.
*/
int yespower_tls(const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst)
const yespower_params_t *params, yespower_binary_t *dst, int thrid )
{
static __thread int initialized = 0;
static __thread yespower_local_t local;
@@ -1128,7 +1139,7 @@ int yespower_tls(const uint8_t *src, size_t srclen,
initialized = 1;
}
return yespower(&local, src, srclen, params, dst);
return yespower( &local, src, srclen, params, dst, thrid );
}
int yespower_init_local(yespower_local_t *local)

9
algo/yespower/yespower.h
View File

@@ -32,6 +32,7 @@
#include <stdint.h>
#include <stdlib.h> /* for size_t */
#include "miner.h"
#ifdef __cplusplus
extern "C" {
@@ -109,11 +110,11 @@ extern int yespower_free_local(yespower_local_t *local);
*/
extern int yespower(yespower_local_t *local,
const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst);
const yespower_params_t *params, yespower_binary_t *dst, int thrid);
extern int yespower_b2b(yespower_local_t *local,
const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst);
const yespower_params_t *params, yespower_binary_t *dst, int thrid );
/**
* yespower_tls(src, srclen, params, dst):
@@ -125,10 +126,10 @@ extern int yespower_b2b(yespower_local_t *local,
* MT-safe as long as dst is local to the thread.
*/
extern int yespower_tls(const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst);
const yespower_params_t *params, yespower_binary_t *dst, int thr_id);
extern int yespower_b2b_tls(const uint8_t *src, size_t srclen,
const yespower_params_t *params, yespower_binary_t *dst);
const yespower_params_t *params, yespower_binary_t *dst, int thr_id);
#ifdef __cplusplus
}