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

v3.9.5.3

Browse Source
This commit is contained in:
Jay D Dee
2019-07-12 10:42:38 -04:00
parent 9abc19a30a
commit e625ed5420
31 changed files with 1269 additions and 1188 deletions
Download Patch File Download Diff File Expand all files Collapse all files

145
algo/shavite/shavite-hash-2way.c
View File

@@ -20,6 +20,7 @@ static const uint32_t IV512[] =
static void
c512_2way( shavite512_2way_context *ctx, const void *msg )
{
const __m128i zero = _mm_setzero_si128();
__m256i p0, p1, p2, p3, x;
__m256i k00, k01, k02, k03, k10, k11, k12, k13;
__m256i *m = (__m256i*)msg;
@@ -33,24 +34,24 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
// round
k00 = m[0];
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k00 ), zero );
k01 = m[1];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = m[2];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = m[3];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p0 = _mm256_xor_si256( p0, x );
k10 = m[4];
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k10 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k10 ), zero );
k11 = m[5];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = m[6];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = m[7];
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p2 = _mm256_xor_si256( p2, x );
@@ -59,129 +60,129 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
// round 1, 5, 9
k00 = _mm256_xor_si256( k13, mm256_ror1x32_128(
mm256_aesenc_2x128( k00 ) ) );
mm256_aesenc_2x128( k00, zero ) ) );
if ( r == 0 )
k00 = _mm256_xor_si256( k00, _mm256_set_epi32(
~ctx->count3, ctx->count2, ctx->count1, ctx->count0,
~ctx->count3, ctx->count2, ctx->count1, ctx->count0 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ), zero );
k01 = _mm256_xor_si256( k00,
mm256_ror1x32_128( mm256_aesenc_2x128( k01 ) ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k01, zero ) ) );
if ( r == 1 )
k01 = _mm256_xor_si256( k01, _mm256_set_epi32(
~ctx->count0, ctx->count1, ctx->count2, ctx->count3,
~ctx->count0, ctx->count1, ctx->count2, ctx->count3 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( k01,
mm256_ror1x32_128( mm256_aesenc_2x128( k02 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k02, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( k02,
mm256_ror1x32_128( mm256_aesenc_2x128( k03 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
mm256_ror1x32_128( 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_ror1x32_128( mm256_aesenc_2x128( k10 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k10, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ), zero );
k11 = _mm256_xor_si256( k10,
mm256_ror1x32_128( mm256_aesenc_2x128( k11 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k11, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( k11,
mm256_ror1x32_128( mm256_aesenc_2x128( k12 ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k12, zero ) ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( k12,
mm256_ror1x32_128( mm256_aesenc_2x128( k13 ) ) );
mm256_ror1x32_128( mm256_aesenc_2x128( k13, zero ) ) );
if ( r == 2 )
k13 = _mm256_xor_si256( k13, _mm256_set_epi32(
~ctx->count1, ctx->count0, ctx->count3, ctx->count2,
~ctx->count1, ctx->count0, ctx->count3, ctx->count2 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p1 = _mm256_xor_si256( p1, x );
// round 2, 6, 10
k00 = _mm256_xor_si256( k00, mm256_ror2x256hi_1x32( k12, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k00 ), zero );
k01 = _mm256_xor_si256( k01, mm256_ror2x256hi_1x32( k13, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( k02, mm256_ror2x256hi_1x32( k00, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( k03, mm256_ror2x256hi_1x32( k01, k02 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p2 = _mm256_xor_si256( p2, x );
k10 = _mm256_xor_si256( k10, mm256_ror2x256hi_1x32( k02, k03 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k10 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k10 ), zero );
k11 = _mm256_xor_si256( k11, mm256_ror2x256hi_1x32( k03, k10 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( k12, mm256_ror2x256hi_1x32( k10, k11 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( k13, mm256_ror2x256hi_1x32( k11, k12 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p0 = _mm256_xor_si256( p0, x );
// round 3, 7, 11
k00 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k00 ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k00 ) );
mm256_aesenc_2x128( k00, zero ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k00 ), zero );
k01 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k01 ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
mm256_aesenc_2x128( k01, zero ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k02 ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
mm256_aesenc_2x128( k02, zero ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k03 ) ), k02 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
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_ror1x32_128(
mm256_aesenc_2x128( k10 ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k10 ) );
mm256_aesenc_2x128( k10, zero ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k10 ), zero );
k11 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k11 ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
mm256_aesenc_2x128( k11, zero ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k12 ) ), k11 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ) );
mm256_aesenc_2x128( k12, zero ) ), k11 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k13 ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
mm256_aesenc_2x128( k13, zero ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p3 = _mm256_xor_si256( p3, x );
// round 4, 8, 12
k00 = _mm256_xor_si256( k00, mm256_ror2x256hi_1x32( k12, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p1, k00 ), zero );
k01 = _mm256_xor_si256( k01, mm256_ror2x256hi_1x32( k13, k00 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( k02, mm256_ror2x256hi_1x32( k00, k01 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( k03, mm256_ror2x256hi_1x32( k01, k02 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ), zero );
p0 = _mm256_xor_si256( p0, x );
k10 = _mm256_xor_si256( k10, mm256_ror2x256hi_1x32( k02, k03 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k10 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( p3, k10 ), zero );
k11 = _mm256_xor_si256( k11, mm256_ror2x256hi_1x32( k03, k10 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = _mm256_xor_si256( k12, mm256_ror2x256hi_1x32( k10, k11 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( k13, mm256_ror2x256hi_1x32( k11, k12 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p2 = _mm256_xor_si256( p2, x );
@@ -190,36 +191,36 @@ c512_2way( shavite512_2way_context *ctx, const void *msg )
// round 13
k00 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k00 ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ) );
mm256_aesenc_2x128( k00, zero ) ), k13 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p0, k00 ), zero );
k01 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k01 ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ) );
mm256_aesenc_2x128( k01, zero ) ), k00 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k01 ), zero );
k02 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k02 ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ) );
mm256_aesenc_2x128( k02, zero ) ), k01 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k02 ), zero );
k03 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k03 ) ), k02 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k03 ) );
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_ror1x32_128(
mm256_aesenc_2x128( k10 ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ) );
mm256_aesenc_2x128( k10, zero ) ), k03 );
x = mm256_aesenc_2x128( _mm256_xor_si256( p2, k10 ), zero );
k11 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k11 ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ) );
mm256_aesenc_2x128( k11, zero ) ), k10 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k11 ), zero );
k12 = mm256_ror1x32_128( mm256_aesenc_2x128( k12 ) );
k12 = mm256_ror1x32_128( 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 ) );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k12 ), zero );
k13 = _mm256_xor_si256( mm256_ror1x32_128(
mm256_aesenc_2x128( k13 ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ) );
mm256_aesenc_2x128( k13, zero ) ), k12 );
x = mm256_aesenc_2x128( _mm256_xor_si256( x, k13 ), zero );
p1 = _mm256_xor_si256( p1, x );

145
algo/shavite/sph-shavite-aesni.c
View File

@@ -87,6 +87,7 @@ static const sph_u32 IV512[] = {
static void
c512( sph_shavite_big_context *sc, const void *msg )
{
const __m128i zero = _mm_setzero_si128();
__m128i p0, p1, p2, p3, x;
__m128i k00, k01, k02, k03, k10, k11, k12, k13;
__m128i *m = (__m128i*)msg;
@@ -101,38 +102,38 @@ c512( sph_shavite_big_context *sc, const void *msg )
// round
k00 = m[0];
x = _mm_xor_si128( p1, k00 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k01 = m[1];
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k02 = m[2];
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k03 = m[3];
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p0 = _mm_xor_si128( p0, x );
k10 = m[4];
x = _mm_xor_si128( p3, k10 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k11 = m[5];
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k12 = m[6];
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k13 = m[7];
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p2 = _mm_xor_si128( p2, x );
for ( r = 0; r < 3; r ++ )
{
// round 1, 5, 9
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
if ( r == 0 )
@@ -140,8 +141,8 @@ c512( sph_shavite_big_context *sc, const void *msg )
~sc->count3, sc->count2, sc->count1, sc->count0 ) );
x = _mm_xor_si128( p0, k00 );
x = _mm_aesenc_si128( x, m128_zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
if ( r == 1 )
@@ -149,32 +150,32 @@ c512( sph_shavite_big_context *sc, const void *msg )
~sc->count0, sc->count1, sc->count2, sc->count3 ) );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p3 = _mm_xor_si128( p3, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p2, k10 );
x = _mm_aesenc_si128( x, m128_zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, zero ) );
k12 = _mm_xor_si128( k12, k11 );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, m128_zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
if ( r == 2 )
@@ -182,78 +183,78 @@ c512( sph_shavite_big_context *sc, const void *msg )
~sc->count1, sc->count0, sc->count3, sc->count2 ) );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p1 = _mm_xor_si128( p1, x );
// round 2, 6, 10
k00 = _mm_xor_si128( k00, mm128_ror256hi_1x32( k12, k13 ) );
x = _mm_xor_si128( p3, k00 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k01 = _mm_xor_si128( k01, mm128_ror256hi_1x32( k13, k00 ) );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k02 = _mm_xor_si128( k02, mm128_ror256hi_1x32( k00, k01 ) );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k03 = _mm_xor_si128( k03, mm128_ror256hi_1x32( k01, k02 ) );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p2 = _mm_xor_si128( p2, x );
k10 = _mm_xor_si128( k10, mm128_ror256hi_1x32( k02, k03 ) );
x = _mm_xor_si128( p1, k10 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k11 = _mm_xor_si128( k11, mm128_ror256hi_1x32( k03, k10 ) );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k12 = _mm_xor_si128( k12, mm128_ror256hi_1x32( k10, k11 ) );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k13 = _mm_xor_si128( k13, mm128_ror256hi_1x32( k11, k12 ) );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p0 = _mm_xor_si128( p0, x );
// round 3, 7, 11
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
x = _mm_xor_si128( p2, k00 );
x = _mm_aesenc_si128( x, m128_zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p1 = _mm_xor_si128( p1, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p0, k10 );
x = _mm_aesenc_si128( x, m128_zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, zero ) );
k12 = _mm_xor_si128( k12, k11 );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, m128_zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p3 = _mm_xor_si128( p3, x );
@@ -261,73 +262,73 @@ c512( sph_shavite_big_context *sc, const void *msg )
k00 = _mm_xor_si128( k00, mm128_ror256hi_1x32( k12, k13 ) );
x = _mm_xor_si128( p1, k00 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k01 = _mm_xor_si128( k01, mm128_ror256hi_1x32( k13, k00 ) );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k02 = _mm_xor_si128( k02, mm128_ror256hi_1x32( k00, k01 ) );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k03 = _mm_xor_si128( k03, mm128_ror256hi_1x32( k01, k02 ) );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p0 = _mm_xor_si128( p0, x );
k10 = _mm_xor_si128( k10, mm128_ror256hi_1x32( k02, k03 ) );
x = _mm_xor_si128( p3, k10 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k11 = _mm_xor_si128( k11, mm128_ror256hi_1x32( k03, k10 ) );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k12 = _mm_xor_si128( k12, mm128_ror256hi_1x32( k10, k11 ) );
x = _mm_xor_si128( x, k12 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
k13 = _mm_xor_si128( k13, mm128_ror256hi_1x32( k11, k12 ) );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p2 = _mm_xor_si128( p2, x );
}
// round 13
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, m128_zero ) );
k00 = mm128_ror_1x32( _mm_aesenc_si128( k00, zero ) );
k00 = _mm_xor_si128( k00, k13 );
x = _mm_xor_si128( p0, k00 );
x = _mm_aesenc_si128( x, m128_zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k01 = mm128_ror_1x32( _mm_aesenc_si128( k01, zero ) );
k01 = _mm_xor_si128( k01, k00 );
x = _mm_xor_si128( x, k01 );
x = _mm_aesenc_si128( x, m128_zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k02 = mm128_ror_1x32( _mm_aesenc_si128( k02, zero ) );
k02 = _mm_xor_si128( k02, k01 );
x = _mm_xor_si128( x, k02 );
x = _mm_aesenc_si128( x, m128_zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k03 = mm128_ror_1x32( _mm_aesenc_si128( k03, zero ) );
k03 = _mm_xor_si128( k03, k02 );
x = _mm_xor_si128( x, k03 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p3 = _mm_xor_si128( p3, x );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, m128_zero ) );
k10 = mm128_ror_1x32( _mm_aesenc_si128( k10, zero ) );
k10 = _mm_xor_si128( k10, k03 );
x = _mm_xor_si128( p2, k10 );
x = _mm_aesenc_si128( x, m128_zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k11 = mm128_ror_1x32( _mm_aesenc_si128( k11, zero ) );
k11 = _mm_xor_si128( k11, k10 );
x = _mm_xor_si128( x, k11 );
x = _mm_aesenc_si128( x, m128_zero );
k12 = mm128_ror_1x32( _mm_aesenc_si128( k12, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k12 = mm128_ror_1x32( _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, m128_zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, m128_zero ) );
x = _mm_aesenc_si128( x, zero );
k13 = mm128_ror_1x32( _mm_aesenc_si128( k13, zero ) );
k13 = _mm_xor_si128( k13, k12 );
x = _mm_xor_si128( x, k13 );
x = _mm_aesenc_si128( x, m128_zero );
x = _mm_aesenc_si128( x, zero );
p1 = _mm_xor_si128( p1, x );