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

v3.7.4

Browse Source
This commit is contained in:
Jay D Dee
2021-09-29 17:31:16 -04:00
parent 9b905fccc8
commit 2cd1507c2e
80 changed files with 8145 additions and 2097 deletions
Download Patch File Download Diff File Expand all files Collapse all files

52
algo/shavite/shavite-hash-2way.c
View File

@@ -20,8 +20,8 @@ static const uint32_t IV512[] =
#define mm256_ror2x256hi_1x32( a, b ) \
_mm256_blend_epi32( mm256_ror128_32( a ), \
mm256_ror128_32( b ), 0x88 )
_mm256_blend_epi32( mm256_shuflr128_32( a ), \
mm256_shuflr128_32( b ), 0x88 )
#if defined(__VAES__)
@@ -78,7 +78,7 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
{
// round 1, 5, 9
k00 = _mm256_xor_si256( k13, mm256_ror128_32(
k00 = _mm256_xor_si256( k13, mm256_shuflr128_32(
mm256_aesenc_2x128( k00, zero ) ) );
if ( r == 0 )
@@ -88,7 +88,7 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ), zero );
k01 = _mm256_xor_si256( k00,
mm256_ror128_32( mm256_aesenc_2x128( k01, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k01, zero ) ) );
if ( r == 1 )
k01 = _mm256_xor_si256( k01, _mm256_set_epi32(
@@ -97,25 +97,25 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( k01,
mm256_ror128_32( mm256_aesenc_2x128( k02, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k02, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( k02,
mm256_ror128_32( mm256_aesenc_2x128( k03, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k03, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p3 = _mm256_xor_si256( p3, x );
k10 = _mm256_xor_si256( k03,
mm256_ror128_32( mm256_aesenc_2x128( k10, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k10, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ), zero );
k11 = _mm256_xor_si256( k10,
mm256_ror128_32( mm256_aesenc_2x128( k11, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k11, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( k11,
mm256_ror128_32( mm256_aesenc_2x128( k12, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k12, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( k12,
mm256_ror128_32( mm256_aesenc_2x128( k13, zero ) ) );
mm256_shuflr128_32( mm256_aesenc_2x128( k13, zero ) ) );
if ( r == 2 )
k13 = _mm256_xor_si256( k13, _mm256_set_epi32(
@@ -151,31 +151,31 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
// round 3, 7, 11
k00 = _mm256_xor_si256( mm256_ror128_32(
k00 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k00, zero ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k00 ), zero );
k01 = _mm256_xor_si256( mm256_ror128_32(
k01 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k01, zero ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( mm256_ror128_32(
k02 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k02, zero ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( mm256_ror128_32(
k03 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k03, zero ) ), k02 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p1 = _mm256_xor_si256( p1, x );
k10 = _mm256_xor_si256( mm256_ror128_32(
k10 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k10, zero ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k10 ), zero );
k11 = _mm256_xor_si256( mm256_ror128_32(
k11 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k11, zero ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( mm256_ror128_32(
k12 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k12, zero ) ), k11 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( mm256_ror128_32(
k13 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k13, zero ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
@@ -209,35 +209,35 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
// round 13
k00 = _mm256_xor_si256( mm256_ror128_32(
k00 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k00, zero ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ), zero );
k01 = _mm256_xor_si256( mm256_ror128_32(
k01 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k01, zero ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( mm256_ror128_32(
k02 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k02, zero ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( mm256_ror128_32(
k03 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k03, zero ) ), k02 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p3 = _mm256_xor_si256( p3, x );
k10 = _mm256_xor_si256( mm256_ror128_32(
k10 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k10, zero ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ), zero );
k11 = _mm256_xor_si256( mm256_ror128_32(
k11 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k11, zero ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = mm256_ror128_32( mm256_aesenc_2x128( k12, zero ) );
k12 = mm256_shuflr128_32( mm256_aesenc_2x128( k12, zero ) );
k12 = _mm256_xor_si256( k12, _mm256_xor_si256( k11, _mm256_set_epi32(
~ctx->count2, ctx->count3, ctx->count0, ctx->count1,
~ctx->count2, ctx->count3, ctx->count0, ctx->count1 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( mm256_ror128_32(
k13 = _mm256_xor_si256( mm256_shuflr128_32(
mm256_aesenc_2x128( k13, zero ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );

54
algo/shavite/shavite-hash-4way.c
View File

@@ -12,8 +12,8 @@ static const uint32_t IV512[] =
};
#define mm512_ror2x512hi_1x32( a, b ) \
_mm512_mask_blend_epi32( 0x8888, mm512_ror128_32( a ), \
mm512_ror128_32( b ) )
_mm512_mask_blend_epi32( 0x8888, mm512_shuflr128_32( a ), \
mm512_shuflr128_32( b ) )
static void
c512_4way( shavite512_4way_context *ctx, const void *msg )
@@ -60,7 +60,7 @@ c512_4way( shavite512_4way_context *ctx, const void *msg )
{
// round 1, 5, 9
K0 = _mm512_xor_si512( K7, mm512_ror128_32(
K0 = _mm512_xor_si512( K7, mm512_shuflr128_32(
_mm512_aesenc_epi128( K0, m512_zero ) ) );
if ( r == 0 )
@@ -69,33 +69,33 @@ c512_4way( shavite512_4way_context *ctx, const void *msg )
X = _mm512_aesenc_epi128( _mm512_xor_si512( P0, K0 ), m512_zero );
K1 = _mm512_xor_si512( K0,
mm512_ror128_32( _mm512_aesenc_epi128( K1, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K1, m512_zero ) ) );
if ( r == 1 )
K1 = _mm512_xor_si512( K1, mm512_ror128_32(
K1 = _mm512_xor_si512( K1, mm512_shuflr128_32(
_mm512_mask_xor_epi32( count, 0x1111, count, m512_neg1 ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K1 ), m512_zero );
K2 = _mm512_xor_si512( K1,
mm512_ror128_32( _mm512_aesenc_epi128( K2, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K2, m512_zero ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K2 ), m512_zero );
K3 = _mm512_xor_si512( K2,
mm512_ror128_32( _mm512_aesenc_epi128( K3, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K3, m512_zero ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K3 ), m512_zero );
P3 = _mm512_xor_si512( P3, X );
K4 = _mm512_xor_si512( K3,
mm512_ror128_32( _mm512_aesenc_epi128( K4, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K4, m512_zero ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( P2, K4 ), m512_zero );
K5 = _mm512_xor_si512( K4,
mm512_ror128_32( _mm512_aesenc_epi128( K5, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K5, m512_zero ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K5 ), m512_zero );
K6 = _mm512_xor_si512( K5,
mm512_ror128_32( _mm512_aesenc_epi128( K6, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K6, m512_zero ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K6 ), m512_zero );
K7 = _mm512_xor_si512( K6,
mm512_ror128_32( _mm512_aesenc_epi128( K7, m512_zero ) ) );
mm512_shuflr128_32( _mm512_aesenc_epi128( K7, m512_zero ) ) );
if ( r == 2 )
K7 = _mm512_xor_si512( K7, mm512_swap128_64(
@@ -130,31 +130,31 @@ c512_4way( shavite512_4way_context *ctx, const void *msg )
// round 3, 7, 11
K0 = _mm512_xor_si512( mm512_ror128_32(
K0 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K0, m512_zero ) ), K7 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( P2, K0 ), m512_zero );
K1 = _mm512_xor_si512( mm512_ror128_32(
K1 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K1, m512_zero ) ), K0 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K1 ), m512_zero );
K2 = _mm512_xor_si512( mm512_ror128_32(
K2 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K2, m512_zero ) ), K1 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K2 ), m512_zero );
K3 = _mm512_xor_si512( mm512_ror128_32(
K3 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K3, m512_zero ) ), K2 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K3 ), m512_zero );
P1 = _mm512_xor_si512( P1, X );
K4 = _mm512_xor_si512( mm512_ror128_32(
K4 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K4, m512_zero ) ), K3 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( P0, K4 ), m512_zero );
K5 = _mm512_xor_si512( mm512_ror128_32(
K5 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K5, m512_zero ) ), K4 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K5 ), m512_zero );
K6 = _mm512_xor_si512( mm512_ror128_32(
K6 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K6, m512_zero ) ), K5 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K6 ), m512_zero );
K7 = _mm512_xor_si512( mm512_ror128_32(
K7 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K7, m512_zero ) ), K6 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K7 ), m512_zero );
@@ -187,34 +187,34 @@ c512_4way( shavite512_4way_context *ctx, const void *msg )
// round 13
K0 = _mm512_xor_si512( mm512_ror128_32(
K0 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K0, m512_zero ) ), K7 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( P0, K0 ), m512_zero );
K1 = _mm512_xor_si512( mm512_ror128_32(
K1 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K1, m512_zero ) ), K0 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K1 ), m512_zero );
K2 = _mm512_xor_si512( mm512_ror128_32(
K2 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K2, m512_zero ) ), K1 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K2 ), m512_zero );
K3 = _mm512_xor_si512( mm512_ror128_32(
K3 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K3, m512_zero ) ), K2 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K3 ), m512_zero );
P3 = _mm512_xor_si512( P3, X );
K4 = _mm512_xor_si512( mm512_ror128_32(
K4 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K4, m512_zero ) ), K3 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( P2, K4 ), m512_zero );
K5 = _mm512_xor_si512( mm512_ror128_32(
K5 = _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K5, m512_zero ) ), K4 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K5 ), m512_zero );
K6 = mm512_ror128_32( _mm512_aesenc_epi128( K6, m512_zero ) );
K6 = mm512_shuflr128_32( _mm512_aesenc_epi128( K6, m512_zero ) );
K6 = _mm512_xor_si512( K6, _mm512_xor_si512( K5, _mm512_set4_epi32(
~ctx->count2, ctx->count3, ctx->count0, ctx->count1 ) ) );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K6 ), m512_zero );
K7= _mm512_xor_si512( mm512_ror128_32(
K7= _mm512_xor_si512( mm512_shuflr128_32(
_mm512_aesenc_epi128( K7, m512_zero ) ), K6 );
X = _mm512_aesenc_epi128( _mm512_xor_si512( X, K7 ), m512_zero );

52
algo/shavite/sph-shavite-aesni.c
View File

@@ -74,15 +74,15 @@ static const sph_u32 IV512[] = {
#endif
/*
#if defined(__AVX2__)
// 2 way version of above
// a[7:0] = { b[4], a[7], a[6], a[5], b[0], a[3], a[2], a[1] }
#define mm256_ror2x256hi_1x32( a, b ) \
_mm256_blend_epi32( mm256_ror256_1x32( a ), \
mm256_rol256_3x32( b ), 0x88 )
#endif
*/
static void
c512( sph_shavite_big_context *sc, const void *msg )
@@ -135,7 +135,7 @@ c512( sph_shavite_big_context *sc, const void *msg )
for ( r = 0; r < 3; r ++ )
{
// round 1, 5, 9
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = mm128_shuflr_32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
if ( r == 0 )
@@ -144,7 +144,7 @@ c512( sph_shavite_big_context *sc, const void *msg )
x = _mm_xor_si128( p0, k00 );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = mm128_shuflr_32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
if ( r == 1 )
@@ -153,31 +153,31 @@ c512( sph_shavite_big_context *sc, const void *msg )
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = mm128_shuflr_32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = mm128_shuflr_32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, zero );
p3 = _mm_xor_si128( p3, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = mm128_shuflr_32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p2, k10 );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = mm128_shuflr_32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, zero ) );
k12 = mm128_shuflr_32( _mm_aesenc_si128( k12, zero ) );
k12 = _mm_xor_si128( k12, k11 );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = mm128_shuflr_32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
if ( r == 2 )
@@ -222,38 +222,38 @@ c512( sph_shavite_big_context *sc, const void *msg )
// round 3, 7, 11
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = mm128_shuflr_32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
x = _mm_xor_si128( p2, k00 );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = mm128_shuflr_32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = mm128_shuflr_32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = mm128_shuflr_32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, zero );
p1 = _mm_xor_si128( p1, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = mm128_shuflr_32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p0, k10 );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = mm128_shuflr_32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, zero ) );
k12 = mm128_shuflr_32( _mm_aesenc_si128( k12, zero ) );
k12 = _mm_xor_si128( k12, k11 );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = mm128_shuflr_32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, zero );
@@ -295,39 +295,39 @@ c512( sph_shavite_big_context *sc, const void *msg )
// round 13
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = mm128_shuflr_32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
x = _mm_xor_si128( p0, k00 );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = mm128_shuflr_32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = mm128_shuflr_32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = mm128_shuflr_32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, zero );
p3 = _mm_xor_si128( p3, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = mm128_shuflr_32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p2, k10 );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = mm128_shuflr_32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, zero ) );
k12 = mm128_shuflr_32( _mm_aesenc_si128( k12, zero ) );
k12 = _mm_xor_si128( k12, _mm_xor_si128( k11, _mm_set_epi32(
~sc->count2, sc->count3, sc->count0, sc->count1 ) ) );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = mm128_shuflr_32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, zero );