Skip to content

Commit

Permalink
v3.23.0
Browse files Browse the repository at this point in the history
  • Loading branch information
@JayDDee
JayDDee committed Aug 31, 2023
1 parent 57a6b7b commit 4378d2f
Show file tree
Hide file tree
Showing 72 changed files with 10,207 additions and 2,205 deletions.
12 changes: 11 additions & 1 deletion RELEASE_NOTES
View file
Original file line number Diff line number Diff line change
Expand Up @@ -65,9 +65,19 @@ If not what makes it happen or not happen?
Change Log
----------

v3.23.0

#398: Prevent GBT fallback to Getwork on network error.
#398: Prevent excessive logs when conditional mining is paused when mining solo.
Fix a false start if stratum doesn't immediately send a new job after connecting.
Tweak diagonal shuffle in Blake2b & Blake256 1-way SIMD to reduce latency.
CPUID support for AVX10.
Initial changes to AVX2 targeted code in preparation for AVX10.
Code cleanup and miscellaneous small improvements.

v3.22.3

Data interleaving and byte swap optimizations iwith AVX2, AVX512 & AVX512VBMI.
Data interleaving and byte swap optimizations with AVX2, AVX512 & AVX512VBMI.
Faster Luffa with AVX2 & AVX512.
Other small optimizations.
Some code cleanup.
Expand Down
4 changes: 2 additions & 2 deletions algo-gate-api.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -171,7 +171,7 @@ int scanhash_4way_64in_32out( struct work *work, uint32_t max_nonce,
}
}
*noncev = _mm256_add_epi32( *noncev,
m256_const1_64( 0x0000000400000000 ) );
_mm256_set1_epi64x( 0x0000000400000000 ) );
n += 4;
} while ( likely( ( n <= last_nonce ) && !work_restart[thr_id].restart ) );
pdata[19] = n;
Expand Down Expand Up @@ -227,7 +227,7 @@ int scanhash_8way_64in_32out( struct work *work, uint32_t max_nonce,
}
}
*noncev = _mm512_add_epi32( *noncev,
m512_const1_64( 0x0000000800000000 ) );
_mm512_set1_epi64( 0x0000000800000000 ) );
n += 8;
} while ( likely( ( n < last_nonce ) && !work_restart[thr_id].restart ) );
pdata[19] = n;
Expand Down
9 changes: 6 additions & 3 deletions algo-gate-api.h
View file
Original file line number Diff line number Diff line change
Expand Up @@ -94,10 +94,13 @@ typedef uint32_t set_t;
#define SSE42_OPT 4
#define AVX_OPT 8 // Sandybridge
#define AVX2_OPT 0x10 // Haswell, Zen1
#define SHA_OPT 0x20 // Zen1, Icelake (sha256)
#define AVX512_OPT 0x40 // Skylake-X (AVX512[F,VL,DQ,BW])
#define VAES_OPT 0x80 // Icelake (VAES & AVX512)
#define SHA_OPT 0x20 // Zen1, Icelake (deprecated)
#define AVX512_OPT 0x40 // Skylake-X, Zen4 (AVX512[F,VL,DQ,BW])
#define VAES_OPT 0x80 // Icelake, Zen3

// AVX10 does not have explicit algo features:
// AVX10_512 is compatible with AVX512 + VAES
// AVX10_256 is compatible with AVX2 + VAES

// return set containing all elements from sets a & b
inline set_t set_union ( set_t a, set_t b ) { return a | b; }
Expand Down
46 changes: 46 additions & 0 deletions algo/blake/blake256-hash-4way.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -308,7 +308,52 @@ static const sph_u32 CS[16] = {
/////////////////////////////////////////
//
// Blake-256 1 way SIMD
// Only used for prehash, otherwise 4way is used with SSE2.

// optimize shuffles to reduce latency caused by dependencies on V1.
#define BLAKE256_ROUND( r ) \
{ \
V0 = _mm_add_epi32( V0, _mm_add_epi32( V1, \
_mm_set_epi32( CSx( r, 7 ) ^ Mx( r, 6 ), \
CSx( r, 5 ) ^ Mx( r, 4 ), \
CSx( r, 3 ) ^ Mx( r, 2 ), \
CSx( r, 1 ) ^ Mx( r, 0 ) ) ) ); \
V3 = mm128_swap32_16( _mm_xor_si128( V3, V0 ) ); \
V2 = _mm_add_epi32( V2, V3 ); \
V1 = mm128_ror_32( _mm_xor_si128( V1, V2 ), 12 ); \
V0 = _mm_add_epi32( V0, _mm_add_epi32( V1, \
_mm_set_epi32( CSx( r, 6 ) ^ Mx( r, 7 ), \
CSx( r, 4 ) ^ Mx( r, 5 ), \
CSx( r, 2 ) ^ Mx( r, 3 ), \
CSx( r, 0 ) ^ Mx( r, 1 ) ) ) ); \
V3 = mm128_shuflr32_8( _mm_xor_si128( V3, V0 ) ); \
V2 = _mm_add_epi32( V2, V3 ); \
V1 = mm128_ror_32( _mm_xor_si128( V1, V2 ), 7 ); \
V0 = mm128_shufll_32( V0 ); \
V3 = mm128_swap_64( V3 ); \
V2 = mm128_shuflr_32( V2 ); \
V0 = _mm_add_epi32( V0, _mm_add_epi32( V1, \
_mm_set_epi32( CSx( r, D ) ^ Mx( r, C ), \
CSx( r, B ) ^ Mx( r, A ), \
CSx( r, 9 ) ^ Mx( r, 8 ), \
CSx( r, F ) ^ Mx( r, E ) ) ) ); \
V3 = mm128_swap32_16( _mm_xor_si128( V3, V0 ) ); \
V2 = _mm_add_epi32( V2, V3 ); \
V1 = mm128_ror_32( _mm_xor_si128( V1, V2 ), 12 ); \
V0 = _mm_add_epi32( V0, _mm_add_epi32( V1, \
_mm_set_epi32( CSx( r, C ) ^ Mx( r, D ), \
CSx( r, A ) ^ Mx( r, B ), \
CSx( r, 8 ) ^ Mx( r, 9 ), \
CSx( r, E ) ^ Mx( r, F ) ) ) ); \
V3 = mm128_shuflr32_8( _mm_xor_si128( V3, V0 ) ); \
V2 = _mm_add_epi32( V2, V3 ); \
V1 = mm128_ror_32( _mm_xor_si128( V1, V2 ), 7 ); \
V0 = mm128_shuflr_32( V0 ); \
V3 = mm128_swap_64( V3 ); \
V2 = mm128_shufll_32( V2 ); \
}

/*
#define BLAKE256_ROUND( r ) \
{ \
V0 = _mm_add_epi32( V0, _mm_add_epi32( V1, \
Expand Down Expand Up @@ -350,6 +395,7 @@ static const sph_u32 CS[16] = {
V2 = mm128_swap_64( V2 ); \
V1 = mm128_shufll_32( V1 ); \
}
*/

void blake256_transform_le( uint32_t *H, const uint32_t *buf,
const uint32_t T0, const uint32_t T1 )
Expand Down
68 changes: 34 additions & 34 deletions algo/blake/blake2b-hash-4way.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -252,14 +252,14 @@ static void blake2b_8way_compress( blake2b_8way_ctx *ctx, int last )
v[ 5] = ctx->h[5];
v[ 6] = ctx->h[6];
v[ 7] = ctx->h[7];
v[ 8] = m512_const1_64( 0x6A09E667F3BCC908 );
v[ 9] = m512_const1_64( 0xBB67AE8584CAA73B );
v[10] = m512_const1_64( 0x3C6EF372FE94F82B );
v[11] = m512_const1_64( 0xA54FF53A5F1D36F1 );
v[12] = m512_const1_64( 0x510E527FADE682D1 );
v[13] = m512_const1_64( 0x9B05688C2B3E6C1F );
v[14] = m512_const1_64( 0x1F83D9ABFB41BD6B );
v[15] = m512_const1_64( 0x5BE0CD19137E2179 );
v[ 8] = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
v[ 9] = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
v[10] = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
v[11] = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
v[12] = _mm512_set1_epi64( 0x510E527FADE682D1 );
v[13] = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
v[14] = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
v[15] = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

v[12] = _mm512_xor_si512( v[12], _mm512_set1_epi64( ctx->t[0] ) );
v[13] = _mm512_xor_si512( v[13], _mm512_set1_epi64( ctx->t[1] ) );
Expand Down Expand Up @@ -310,16 +310,16 @@ int blake2b_8way_init( blake2b_8way_ctx *ctx )
{
size_t i;

ctx->h[0] = m512_const1_64( 0x6A09E667F3BCC908 );
ctx->h[1] = m512_const1_64( 0xBB67AE8584CAA73B );
ctx->h[2] = m512_const1_64( 0x3C6EF372FE94F82B );
ctx->h[3] = m512_const1_64( 0xA54FF53A5F1D36F1 );
ctx->h[4] = m512_const1_64( 0x510E527FADE682D1 );
ctx->h[5] = m512_const1_64( 0x9B05688C2B3E6C1F );
ctx->h[6] = m512_const1_64( 0x1F83D9ABFB41BD6B );
ctx->h[7] = m512_const1_64( 0x5BE0CD19137E2179 );
ctx->h[0] = _mm512_set1_epi64( 0x6A09E667F3BCC908 );
ctx->h[1] = _mm512_set1_epi64( 0xBB67AE8584CAA73B );
ctx->h[2] = _mm512_set1_epi64( 0x3C6EF372FE94F82B );
ctx->h[3] = _mm512_set1_epi64( 0xA54FF53A5F1D36F1 );
ctx->h[4] = _mm512_set1_epi64( 0x510E527FADE682D1 );
ctx->h[5] = _mm512_set1_epi64( 0x9B05688C2B3E6C1F );
ctx->h[6] = _mm512_set1_epi64( 0x1F83D9ABFB41BD6B );
ctx->h[7] = _mm512_set1_epi64( 0x5BE0CD19137E2179 );

ctx->h[0] = _mm512_xor_si512( ctx->h[0], m512_const1_64( 0x01010020 ) );
ctx->h[0] = _mm512_xor_si512( ctx->h[0], _mm512_set1_epi64( 0x01010020 ) );

ctx->t[0] = 0;
ctx->t[1] = 0;
Expand Down Expand Up @@ -419,14 +419,14 @@ static void blake2b_4way_compress( blake2b_4way_ctx *ctx, int last )
v[ 5] = ctx->h[5];
v[ 6] = ctx->h[6];
v[ 7] = ctx->h[7];
v[ 8] = m256_const1_64( 0x6A09E667F3BCC908 );
v[ 9] = m256_const1_64( 0xBB67AE8584CAA73B );
v[10] = m256_const1_64( 0x3C6EF372FE94F82B );
v[11] = m256_const1_64( 0xA54FF53A5F1D36F1 );
v[12] = m256_const1_64( 0x510E527FADE682D1 );
v[13] = m256_const1_64( 0x9B05688C2B3E6C1F );
v[14] = m256_const1_64( 0x1F83D9ABFB41BD6B );
v[15] = m256_const1_64( 0x5BE0CD19137E2179 );
v[ 8] = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
v[ 9] = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
v[10] = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
v[11] = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
v[12] = _mm256_set1_epi64x( 0x510E527FADE682D1 );
v[13] = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
v[14] = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
v[15] = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

v[12] = _mm256_xor_si256( v[12], _mm256_set1_epi64x( ctx->t[0] ) );
v[13] = _mm256_xor_si256( v[13], _mm256_set1_epi64x( ctx->t[1] ) );
Expand Down Expand Up @@ -477,16 +477,16 @@ int blake2b_4way_init( blake2b_4way_ctx *ctx )
{
size_t i;

ctx->h[0] = m256_const1_64( 0x6A09E667F3BCC908 );
ctx->h[1] = m256_const1_64( 0xBB67AE8584CAA73B );
ctx->h[2] = m256_const1_64( 0x3C6EF372FE94F82B );
ctx->h[3] = m256_const1_64( 0xA54FF53A5F1D36F1 );
ctx->h[4] = m256_const1_64( 0x510E527FADE682D1 );
ctx->h[5] = m256_const1_64( 0x9B05688C2B3E6C1F );
ctx->h[6] = m256_const1_64( 0x1F83D9ABFB41BD6B );
ctx->h[7] = m256_const1_64( 0x5BE0CD19137E2179 );
ctx->h[0] = _mm256_set1_epi64x( 0x6A09E667F3BCC908 );
ctx->h[1] = _mm256_set1_epi64x( 0xBB67AE8584CAA73B );
ctx->h[2] = _mm256_set1_epi64x( 0x3C6EF372FE94F82B );
ctx->h[3] = _mm256_set1_epi64x( 0xA54FF53A5F1D36F1 );
ctx->h[4] = _mm256_set1_epi64x( 0x510E527FADE682D1 );
ctx->h[5] = _mm256_set1_epi64x( 0x9B05688C2B3E6C1F );
ctx->h[6] = _mm256_set1_epi64x( 0x1F83D9ABFB41BD6B );
ctx->h[7] = _mm256_set1_epi64x( 0x5BE0CD19137E2179 );

ctx->h[0] = _mm256_xor_si256( ctx->h[0], m256_const1_64( 0x01010020 ) );
ctx->h[0] = _mm256_xor_si256( ctx->h[0], _mm256_set1_epi64x( 0x01010020 ) );

ctx->t[0] = 0;
ctx->t[1] = 0;
Expand Down
96 changes: 48 additions & 48 deletions algo/blake/blake2s-hash-4way.c
View file
Original file line number Diff line number Diff line change
Expand Up @@ -62,14 +62,14 @@ int blake2s_4way_init( blake2s_4way_state *S, const uint8_t outlen )

memset( S, 0, sizeof( blake2s_4way_state ) );

S->h[0] = m128_const1_64( 0x6A09E6676A09E667ULL );
S->h[1] = m128_const1_64( 0xBB67AE85BB67AE85ULL );
S->h[2] = m128_const1_64( 0x3C6EF3723C6EF372ULL );
S->h[3] = m128_const1_64( 0xA54FF53AA54FF53AULL );
S->h[4] = m128_const1_64( 0x510E527F510E527FULL );
S->h[5] = m128_const1_64( 0x9B05688C9B05688CULL );
S->h[6] = m128_const1_64( 0x1F83D9AB1F83D9ABULL );
S->h[7] = m128_const1_64( 0x5BE0CD195BE0CD19ULL );
S->h[0] = _mm_set1_epi64x( 0x6A09E6676A09E667ULL );
S->h[1] = _mm_set1_epi64x( 0xBB67AE85BB67AE85ULL );
S->h[2] = _mm_set1_epi64x( 0x3C6EF3723C6EF372ULL );
S->h[3] = _mm_set1_epi64x( 0xA54FF53AA54FF53AULL );
S->h[4] = _mm_set1_epi64x( 0x510E527F510E527FULL );
S->h[5] = _mm_set1_epi64x( 0x9B05688C9B05688CULL );
S->h[6] = _mm_set1_epi64x( 0x1F83D9AB1F83D9ABULL );
S->h[7] = _mm_set1_epi64x( 0x5BE0CD195BE0CD19ULL );

// for( int i = 0; i < 8; ++i )
// S->h[i] = _mm_set1_epi32( blake2s_IV[i] );
Expand All @@ -90,18 +90,18 @@ int blake2s_4way_compress( blake2s_4way_state *S, const __m128i* block )
memcpy_128( m, block, 16 );
memcpy_128( v, S->h, 8 );

v[ 8] = m128_const1_64( 0x6A09E6676A09E667ULL );
v[ 9] = m128_const1_64( 0xBB67AE85BB67AE85ULL );
v[10] = m128_const1_64( 0x3C6EF3723C6EF372ULL );
v[11] = m128_const1_64( 0xA54FF53AA54FF53AULL );
v[ 8] = _mm_set1_epi64x( 0x6A09E6676A09E667ULL );
v[ 9] = _mm_set1_epi64x( 0xBB67AE85BB67AE85ULL );
v[10] = _mm_set1_epi64x( 0x3C6EF3723C6EF372ULL );
v[11] = _mm_set1_epi64x( 0xA54FF53AA54FF53AULL );
v[12] = _mm_xor_si128( _mm_set1_epi32( S->t[0] ),
m128_const1_64( 0x510E527F510E527FULL ) );
_mm_set1_epi64x( 0x510E527F510E527FULL ) );
v[13] = _mm_xor_si128( _mm_set1_epi32( S->t[1] ),
m128_const1_64( 0x9B05688C9B05688CULL ) );
_mm_set1_epi64x( 0x9B05688C9B05688CULL ) );
v[14] = _mm_xor_si128( _mm_set1_epi32( S->f[0] ),
m128_const1_64( 0x1F83D9AB1F83D9ABULL ) );
_mm_set1_epi64x( 0x1F83D9AB1F83D9ABULL ) );
v[15] = _mm_xor_si128( _mm_set1_epi32( S->f[1] ),
m128_const1_64( 0x5BE0CD195BE0CD19ULL ) );
_mm_set1_epi64x( 0x5BE0CD195BE0CD19ULL ) );

#define G4W( sigma0, sigma1, a, b, c, d ) \
do { \
Expand Down Expand Up @@ -269,21 +269,21 @@ int blake2s_8way_compress( blake2s_8way_state *S, const __m256i *block )
memcpy_256( m, block, 16 );
memcpy_256( v, S->h, 8 );

v[ 8] = m256_const1_64( 0x6A09E6676A09E667ULL );
v[ 9] = m256_const1_64( 0xBB67AE85BB67AE85ULL );
v[10] = m256_const1_64( 0x3C6EF3723C6EF372ULL );
v[11] = m256_const1_64( 0xA54FF53AA54FF53AULL );
v[ 8] = _mm256_set1_epi64x( 0x6A09E6676A09E667ULL );
v[ 9] = _mm256_set1_epi64x( 0xBB67AE85BB67AE85ULL );
v[10] = _mm256_set1_epi64x( 0x3C6EF3723C6EF372ULL );
v[11] = _mm256_set1_epi64x( 0xA54FF53AA54FF53AULL );
v[12] = _mm256_xor_si256( _mm256_set1_epi32( S->t[0] ),
m256_const1_64( 0x510E527F510E527FULL ) );
_mm256_set1_epi64x( 0x510E527F510E527FULL ) );

v[13] = _mm256_xor_si256( _mm256_set1_epi32( S->t[1] ),
m256_const1_64( 0x9B05688C9B05688CULL ) );
_mm256_set1_epi64x( 0x9B05688C9B05688CULL ) );

v[14] = _mm256_xor_si256( _mm256_set1_epi32( S->f[0] ),
m256_const1_64( 0x1F83D9AB1F83D9ABULL ) );
_mm256_set1_epi64x( 0x1F83D9AB1F83D9ABULL ) );

v[15] = _mm256_xor_si256( _mm256_set1_epi32( S->f[1] ),
m256_const1_64( 0x5BE0CD195BE0CD19ULL ) );
_mm256_set1_epi64x( 0x5BE0CD195BE0CD19ULL ) );

/*
v[ 8] = _mm256_set1_epi32( blake2s_IV[0] );
Expand Down Expand Up @@ -391,14 +391,14 @@ int blake2s_8way_init( blake2s_8way_state *S, const uint8_t outlen )
memset( P->personal, 0, sizeof( P->personal ) );

memset( S, 0, sizeof( blake2s_8way_state ) );
S->h[0] = m256_const1_64( 0x6A09E6676A09E667ULL );
S->h[1] = m256_const1_64( 0xBB67AE85BB67AE85ULL );
S->h[2] = m256_const1_64( 0x3C6EF3723C6EF372ULL );
S->h[3] = m256_const1_64( 0xA54FF53AA54FF53AULL );
S->h[4] = m256_const1_64( 0x510E527F510E527FULL );
S->h[5] = m256_const1_64( 0x9B05688C9B05688CULL );
S->h[6] = m256_const1_64( 0x1F83D9AB1F83D9ABULL );
S->h[7] = m256_const1_64( 0x5BE0CD195BE0CD19ULL );
S->h[0] = _mm256_set1_epi64x( 0x6A09E6676A09E667ULL );
S->h[1] = _mm256_set1_epi64x( 0xBB67AE85BB67AE85ULL );
S->h[2] = _mm256_set1_epi64x( 0x3C6EF3723C6EF372ULL );
S->h[3] = _mm256_set1_epi64x( 0xA54FF53AA54FF53AULL );
S->h[4] = _mm256_set1_epi64x( 0x510E527F510E527FULL );
S->h[5] = _mm256_set1_epi64x( 0x9B05688C9B05688CULL );
S->h[6] = _mm256_set1_epi64x( 0x1F83D9AB1F83D9ABULL );
S->h[7] = _mm256_set1_epi64x( 0x5BE0CD195BE0CD19ULL );


// for( int i = 0; i < 8; ++i )
Expand Down Expand Up @@ -510,21 +510,21 @@ int blake2s_16way_compress( blake2s_16way_state *S, const __m512i *block )
memcpy_512( m, block, 16 );
memcpy_512( v, S->h, 8 );

v[ 8] = m512_const1_64( 0x6A09E6676A09E667ULL );
v[ 9] = m512_const1_64( 0xBB67AE85BB67AE85ULL );
v[10] = m512_const1_64( 0x3C6EF3723C6EF372ULL );
v[11] = m512_const1_64( 0xA54FF53AA54FF53AULL );
v[ 8] = _mm512_set1_epi64( 0x6A09E6676A09E667ULL );
v[ 9] = _mm512_set1_epi64( 0xBB67AE85BB67AE85ULL );
v[10] = _mm512_set1_epi64( 0x3C6EF3723C6EF372ULL );
v[11] = _mm512_set1_epi64( 0xA54FF53AA54FF53AULL );
v[12] = _mm512_xor_si512( _mm512_set1_epi32( S->t[0] ),
m512_const1_64( 0x510E527F510E527FULL ) );
_mm512_set1_epi64( 0x510E527F510E527FULL ) );

v[13] = _mm512_xor_si512( _mm512_set1_epi32( S->t[1] ),
m512_const1_64( 0x9B05688C9B05688CULL ) );
_mm512_set1_epi64( 0x9B05688C9B05688CULL ) );

v[14] = _mm512_xor_si512( _mm512_set1_epi32( S->f[0] ),
m512_const1_64( 0x1F83D9AB1F83D9ABULL ) );
_mm512_set1_epi64( 0x1F83D9AB1F83D9ABULL ) );

v[15] = _mm512_xor_si512( _mm512_set1_epi32( S->f[1] ),
m512_const1_64( 0x5BE0CD195BE0CD19ULL ) );
_mm512_set1_epi64( 0x5BE0CD195BE0CD19ULL ) );


#define G16W( sigma0, sigma1, a, b, c, d) \
Expand Down Expand Up @@ -589,14 +589,14 @@ int blake2s_16way_init( blake2s_16way_state *S, const uint8_t outlen )
memset( P->personal, 0, sizeof( P->personal ) );

memset( S, 0, sizeof( blake2s_16way_state ) );
S->h[0] = m512_const1_64( 0x6A09E6676A09E667ULL );
S->h[1] = m512_const1_64( 0xBB67AE85BB67AE85ULL );
S->h[2] = m512_const1_64( 0x3C6EF3723C6EF372ULL );
S->h[3] = m512_const1_64( 0xA54FF53AA54FF53AULL );
S->h[4] = m512_const1_64( 0x510E527F510E527FULL );
S->h[5] = m512_const1_64( 0x9B05688C9B05688CULL );
S->h[6] = m512_const1_64( 0x1F83D9AB1F83D9ABULL );
S->h[7] = m512_const1_64( 0x5BE0CD195BE0CD19ULL );
S->h[0] = _mm512_set1_epi64( 0x6A09E6676A09E667ULL );
S->h[1] = _mm512_set1_epi64( 0xBB67AE85BB67AE85ULL );
S->h[2] = _mm512_set1_epi64( 0x3C6EF3723C6EF372ULL );
S->h[3] = _mm512_set1_epi64( 0xA54FF53AA54FF53AULL );
S->h[4] = _mm512_set1_epi64( 0x510E527F510E527FULL );
S->h[5] = _mm512_set1_epi64( 0x9B05688C9B05688CULL );
S->h[6] = _mm512_set1_epi64( 0x1F83D9AB1F83D9ABULL );
S->h[7] = _mm512_set1_epi64( 0x5BE0CD195BE0CD19ULL );

uint32_t *p = ( uint32_t * )( P );

Expand Down
Loading

0 comments on commit 4378d2f

Please sign in to comment.