LeoVasanko/RandQuik
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Arrayized the variables, no change in performance but a lot less code.

Browse Source
This commit is contained in:
LeoVasanko
2023-10-28 15:23:30 +00:00
parent 771212029f
commit a6daac4ce6
1 changed files with 181 additions and 201 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/u8-stream.h
+181 -201
View File
@@ -1,38 +1,39 @@
#define VEC8_ROT(A, IMM) \
#define VEC8_ROT(A, IMM) \
_mm256_or_si256(_mm256_slli_epi32(A, IMM), _mm256_srli_epi32(A, (32 - IMM)))
/* same, but replace 2 of the shift/shift/or "rotation" by byte shuffles (8 &
* 16) (better) */
#define VEC8_QUARTERROUND(A, B, C, D) \
x_##A = _mm256_add_epi32(x_##A, x_##B); \
t_##A = _mm256_xor_si256(x_##D, x_##A); \
x_##D = _mm256_shuffle_epi8(t_##A, rot16); \
x_##C = _mm256_add_epi32(x_##C, x_##D); \
t_##C = _mm256_xor_si256(x_##B, x_##C); \
x_##B = VEC8_ROT(t_##C, 12); \
x_##A = _mm256_add_epi32(x_##A, x_##B); \
t_##A = _mm256_xor_si256(x_##D, x_##A); \
x_##D = _mm256_shuffle_epi8(t_##A, rot8); \
x_##C = _mm256_add_epi32(x_##C, x_##D); \
t_##C = _mm256_xor_si256(x_##B, x_##C); \
x_##B = VEC8_ROT(t_##C, 7)
#define VEC8_QUARTERROUND(A, B, C, D) \
x[A] = _mm256_add_epi32(x[A], x[B]); \
t[A] = _mm256_xor_si256(x[D], x[A]); \
x[D] = _mm256_shuffle_epi8(t[A], rot16); \
x[C] = _mm256_add_epi32(x[C], x[D]); \
t[C] = _mm256_xor_si256(x[B], x[C]); \
x[B] = VEC8_ROT(t[C], 12); \
x[A] = _mm256_add_epi32(x[A], x[B]); \
t[A] = _mm256_xor_si256(x[D], x[A]); \
x[D] = _mm256_shuffle_epi8(t[A], rot8); \
x[C] = _mm256_add_epi32(x[C], x[D]); \
t[C] = _mm256_xor_si256(x[B], x[C]); \
x[B] = VEC8_ROT(t[C], 7)
#define VEC8_LINE1(A, B, C, D) \
x_##A = _mm256_add_epi32(x_##A, x_##B); \
x_##D = _mm256_shuffle_epi8(_mm256_xor_si256(x_##D, x_##A), rot16)
#define VEC8_LINE2(A, B, C, D) \
x_##C = _mm256_add_epi32(x_##C, x_##D); \
x_##B = VEC8_ROT(_mm256_xor_si256(x_##B, x_##C), 12)
#define VEC8_LINE3(A, B, C, D) \
x_##A = _mm256_add_epi32(x_##A, x_##B); \
x_##D = _mm256_shuffle_epi8(_mm256_xor_si256(x_##D, x_##A), rot8)
#define VEC8_LINE4(A, B, C, D) \
x_##C = _mm256_add_epi32(x_##C, x_##D); \
x_##B = VEC8_ROT(_mm256_xor_si256(x_##B, x_##C), 7)
#define VEC8_LINE1(A, B, C, D) \
x[A] = _mm256_add_epi32(x[A], x[B]); \
x[D] = _mm256_shuffle_epi8(_mm256_xor_si256(x[D], x[A]), rot16)
#define VEC8_LINE2(A, B, C, D) \
x[C] = _mm256_add_epi32(x[C], x[D]); \
x[B] = VEC8_ROT(_mm256_xor_si256(x[B], x[C]), 12)
#define VEC8_LINE3(A, B, C, D) \
x[A] = _mm256_add_epi32(x[A], x[B]); \
x[D] = _mm256_shuffle_epi8(_mm256_xor_si256(x[D], x[A]), rot8)
#define VEC8_LINE4(A, B, C, D) \
x[C] = _mm256_add_epi32(x[C], x[D]); \
x[B] = VEC8_ROT(_mm256_xor_si256(x[B], x[C]), 7)
#define VEC8_ROUND_SEQ(A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, \
C4, D4) \
#define VEC8_ROUND_SEQ( \
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, D4 \
) \
VEC8_LINE1(A1, B1, C1, D1); \
VEC8_LINE1(A2, B2, C2, D2); \
VEC8_LINE1(A3, B3, C3, D3); \
@@ -50,27 +51,29 @@
VEC8_LINE4(A3, B3, C3, D3); \
VEC8_LINE4(A4, B4, C4, D4)
#define VEC8_ROUND_HALF(A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, \
B4, C4, D4) \
VEC8_LINE1(A1, B1, C1, D1); \
VEC8_LINE1(A2, B2, C2, D2); \
VEC8_LINE2(A1, B1, C1, D1); \
VEC8_LINE2(A2, B2, C2, D2); \
VEC8_LINE3(A1, B1, C1, D1); \
VEC8_LINE3(A2, B2, C2, D2); \
VEC8_LINE4(A1, B1, C1, D1); \
VEC8_LINE4(A2, B2, C2, D2); \
VEC8_LINE1(A3, B3, C3, D3); \
VEC8_LINE1(A4, B4, C4, D4); \
VEC8_LINE2(A3, B3, C3, D3); \
VEC8_LINE2(A4, B4, C4, D4); \
VEC8_LINE3(A3, B3, C3, D3); \
VEC8_LINE3(A4, B4, C4, D4); \
VEC8_LINE4(A3, B3, C3, D3); \
#define VEC8_ROUND_HALF( \
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, D4 \
) \
VEC8_LINE1(A1, B1, C1, D1); \
VEC8_LINE1(A2, B2, C2, D2); \
VEC8_LINE2(A1, B1, C1, D1); \
VEC8_LINE2(A2, B2, C2, D2); \
VEC8_LINE3(A1, B1, C1, D1); \
VEC8_LINE3(A2, B2, C2, D2); \
VEC8_LINE4(A1, B1, C1, D1); \
VEC8_LINE4(A2, B2, C2, D2); \
VEC8_LINE1(A3, B3, C3, D3); \
VEC8_LINE1(A4, B4, C4, D4); \
VEC8_LINE2(A3, B3, C3, D3); \
VEC8_LINE2(A4, B4, C4, D4); \
VEC8_LINE3(A3, B3, C3, D3); \
VEC8_LINE3(A4, B4, C4, D4); \
VEC8_LINE4(A3, B3, C3, D3); \
VEC8_LINE4(A4, B4, C4, D4)
#define VEC8_ROUND_HALFANDHALF(A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, \
A4, B4, C4, D4) \
#define VEC8_ROUND_HALFANDHALF( \
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, D4 \
) \
VEC8_LINE1(A1, B1, C1, D1); \
VEC8_LINE1(A2, B2, C2, D2); \
VEC8_LINE2(A1, B1, C1, D1); \
@@ -88,175 +91,152 @@
VEC8_LINE4(A3, B3, C3, D3); \
VEC8_LINE4(A4, B4, C4, D4)
#define VEC8_ROUND(A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, \
D4) \
VEC8_ROUND_SEQ(A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, \
D4)
#define VEC8_ROUND( \
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, D4 \
) \
VEC8_ROUND_SEQ( \
A1, B1, C1, D1, A2, B2, C2, D2, A3, B3, C3, D3, A4, B4, C4, D4 \
)
if (bytes >= 512)
{
/* constant for shuffling bytes (replacing multiple-of-8 rotates) */
__m256i rot16 =
_mm256_set_epi8(13, 12, 15, 14, 9, 8, 11, 10, 5, 4, 7, 6, 1, 0, 3, 2,
13, 12, 15, 14, 9, 8, 11, 10, 5, 4, 7, 6, 1, 0, 3, 2);
__m256i rot8 =
_mm256_set_epi8(14, 13, 12, 15, 10, 9, 8, 11, 6, 5, 4, 7, 2, 1, 0, 3,
14, 13, 12, 15, 10, 9, 8, 11, 6, 5, 4, 7, 2, 1, 0, 3);
/* the naive way seems as fast (if not a bit faster) than the vector way */
__m256i x_0 = _mm256_set1_epi32(x[0]);
__m256i x_1 = _mm256_set1_epi32(x[1]);
__m256i x_2 = _mm256_set1_epi32(x[2]);
__m256i x_3 = _mm256_set1_epi32(x[3]);
__m256i x_4 = _mm256_set1_epi32(x[4]);
__m256i x_5 = _mm256_set1_epi32(x[5]);
__m256i x_6 = _mm256_set1_epi32(x[6]);
__m256i x_7 = _mm256_set1_epi32(x[7]);
__m256i x_8 = _mm256_set1_epi32(x[8]);
__m256i x_9 = _mm256_set1_epi32(x[9]);
__m256i x_10 = _mm256_set1_epi32(x[10]);
__m256i x_11 = _mm256_set1_epi32(x[11]);
__m256i x_12;
__m256i x_13;
__m256i x_14 = _mm256_set1_epi32(x[14]);
__m256i x_15 = _mm256_set1_epi32(x[15]);
__m256i orig0 = x_0;
__m256i orig1 = x_1;
__m256i orig2 = x_2;
__m256i orig3 = x_3;
__m256i orig4 = x_4;
__m256i orig5 = x_5;
__m256i orig6 = x_6;
__m256i orig7 = x_7;
__m256i orig8 = x_8;
__m256i orig9 = x_9;
__m256i orig10 = x_10;
__m256i orig11 = x_11;
__m256i orig12;
__m256i orig13;
__m256i orig14 = x_14;
__m256i orig15 = x_15;
__m256i t_0, t_1, t_2, t_3, t_4, t_5, t_6, t_7, t_8, t_9, t_10, t_11, t_12,
t_13, t_14, t_15;
const __m256i addv12 = _mm256_set_epi64x(3, 2, 1, 0);
const __m256i addv13 = _mm256_set_epi64x(7, 6, 5, 4);
while (bytes >= 512)
{
__m256i t12, t13;
x_0 = orig0;
x_1 = orig1;
x_2 = orig2;
x_3 = orig3;
x_4 = orig4;
x_5 = orig5;
x_6 = orig6;
x_7 = orig7;
x_8 = orig8;
x_9 = orig9;
x_10 = orig10;
x_11 = orig11;
x_14 = orig14;
x_15 = orig15;
// Calculate the eight parallel counters on x_12 and x_13
uint64_t *counter = (uint64_t *)(x + 12);
t13 = _mm256_broadcastq_epi64(_mm_cvtsi64_si128(*counter));
t12 = _mm256_add_epi64(addv12, t13);
t13 = _mm256_add_epi64(addv13, t13);
x_12 = _mm256_unpacklo_epi32(t12, t13);
x_13 = _mm256_unpackhi_epi32(t12, t13);
t12 = _mm256_unpacklo_epi32(x_12, x_13);
t13 = _mm256_unpackhi_epi32(x_12, x_13);
/* required because unpack* are intra-lane */
const __m256i permute = _mm256_set_epi32(7, 6, 3, 2, 5, 4, 1, 0);
x_12 = _mm256_permutevar8x32_epi32(t12, permute);
x_13 = _mm256_permutevar8x32_epi32(t13, permute);
orig12 = x_12;
orig13 = x_13;
for (int i = 0; i < 10; ++i)
{
VEC8_ROUND(0, 4, 8, 12, 1, 5, 9, 13, 2, 6, 10, 14, 3, 7, 11, 15);
VEC8_ROUND(0, 5, 10, 15, 1, 6, 11, 12, 2, 7, 8, 13, 3, 4, 9, 14);
}
#define ONEQUAD_TRANSPOSE(A, B, C, D) \
{ \
__m128i t0, t1, t2, t3; \
x_##A = _mm256_add_epi32(x_##A, orig##A); \
x_##B = _mm256_add_epi32(x_##B, orig##B); \
x_##C = _mm256_add_epi32(x_##C, orig##C); \
x_##D = _mm256_add_epi32(x_##D, orig##D); \
t_##A = _mm256_unpacklo_epi32(x_##A, x_##B); \
t_##B = _mm256_unpacklo_epi32(x_##C, x_##D); \
t_##C = _mm256_unpackhi_epi32(x_##A, x_##B); \
t_##D = _mm256_unpackhi_epi32(x_##C, x_##D); \
x_##A = _mm256_unpacklo_epi64(t_##A, t_##B); \
x_##B = _mm256_unpackhi_epi64(t_##A, t_##B); \
x_##C = _mm256_unpacklo_epi64(t_##C, t_##D); \
x_##D = _mm256_unpackhi_epi64(t_##C, t_##D); \
_mm_storeu_si128((__m128i *)(c + 0), _mm256_extracti128_si256(x_##A, 0)); \
_mm_storeu_si128((__m128i *)(c + 64), _mm256_extracti128_si256(x_##B, 0)); \
_mm_storeu_si128((__m128i *)(c + 128), _mm256_extracti128_si256(x_##C, 0)); \
_mm_storeu_si128((__m128i *)(c + 192), _mm256_extracti128_si256(x_##D, 0)); \
_mm_storeu_si128((__m128i *)(c + 256), _mm256_extracti128_si256(x_##A, 1)); \
_mm_storeu_si128((__m128i *)(c + 320), _mm256_extracti128_si256(x_##B, 1)); \
_mm_storeu_si128((__m128i *)(c + 384), _mm256_extracti128_si256(x_##C, 1)); \
_mm_storeu_si128((__m128i *)(c + 448), _mm256_extracti128_si256(x_##D, 1)); \
#define ONEQUAD_TRANSPOSE(A, B, C, D) \
{ \
__m128i t0, t1, t2, t3; \
x[A] = _mm256_add_epi32(x[A], orig[A]); \
x[B] = _mm256_add_epi32(x[B], orig[B]); \
x[C] = _mm256_add_epi32(x[C], orig[C]); \
x[D] = _mm256_add_epi32(x[D], orig[D]); \
t[A] = _mm256_unpacklo_epi32(x[A], x[B]); \
t[B] = _mm256_unpacklo_epi32(x[C], x[D]); \
t[C] = _mm256_unpackhi_epi32(x[A], x[B]); \
t[D] = _mm256_unpackhi_epi32(x[C], x[D]); \
x[A] = _mm256_unpacklo_epi64(t[A], t[B]); \
x[B] = _mm256_unpackhi_epi64(t[A], t[B]); \
x[C] = _mm256_unpacklo_epi64(t[C], t[D]); \
x[D] = _mm256_unpackhi_epi64(t[C], t[D]); \
_mm_storeu_si128( \
(__m128i*)(c + 0), _mm256_extracti128_si256(x[A], 0) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 64), _mm256_extracti128_si256(x[B], 0) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 128), _mm256_extracti128_si256(x[C], 0) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 192), _mm256_extracti128_si256(x[D], 0) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 256), _mm256_extracti128_si256(x[A], 1) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 320), _mm256_extracti128_si256(x[B], 1) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 384), _mm256_extracti128_si256(x[C], 1) \
); \
_mm_storeu_si128( \
(__m128i*)(c + 448), _mm256_extracti128_si256(x[D], 1) \
); \
}
#define ONEQUAD(A, B, C, D) ONEQUAD_TRANSPOSE(A, B, C, D)
#define ONEQUAD_UNPCK(A, B, C, D) \
{ \
x_##A = _mm256_add_epi32(x_##A, orig##A); \
x_##B = _mm256_add_epi32(x_##B, orig##B); \
x_##C = _mm256_add_epi32(x_##C, orig##C); \
x_##D = _mm256_add_epi32(x_##D, orig##D); \
t_##A = _mm256_unpacklo_epi32(x_##A, x_##B); \
t_##B = _mm256_unpacklo_epi32(x_##C, x_##D); \
t_##C = _mm256_unpackhi_epi32(x_##A, x_##B); \
t_##D = _mm256_unpackhi_epi32(x_##C, x_##D); \
x_##A = _mm256_unpacklo_epi64(t_##A, t_##B); \
x_##B = _mm256_unpackhi_epi64(t_##A, t_##B); \
x_##C = _mm256_unpacklo_epi64(t_##C, t_##D); \
x_##D = _mm256_unpackhi_epi64(t_##C, t_##D); \
#define ONEQUAD_UNPCK(A, B, C, D) \
{ \
x[A] = _mm256_add_epi32(x[A], orig[A]); \
x[B] = _mm256_add_epi32(x[B], orig[B]); \
x[C] = _mm256_add_epi32(x[C], orig[C]); \
x[D] = _mm256_add_epi32(x[D], orig[D]); \
t[A] = _mm256_unpacklo_epi32(x[A], x[B]); \
t[B] = _mm256_unpacklo_epi32(x[C], x[D]); \
t[C] = _mm256_unpackhi_epi32(x[A], x[B]); \
t[D] = _mm256_unpackhi_epi32(x[C], x[D]); \
x[A] = _mm256_unpacklo_epi64(t[A], t[B]); \
x[B] = _mm256_unpackhi_epi64(t[A], t[B]); \
x[C] = _mm256_unpacklo_epi64(t[C], t[D]); \
x[D] = _mm256_unpackhi_epi64(t[C], t[D]); \
}
#define ONEOCTO(A, B, C, D, A2, B2, C2, D2, c) \
{ \
ONEQUAD_UNPCK(A, B, C, D); \
ONEQUAD_UNPCK(A2, B2, C2, D2); \
t_##A = _mm256_permute2x128_si256(x_##A, x_##A2, 0x20); \
t_##A2 = _mm256_permute2x128_si256(x_##A, x_##A2, 0x31); \
t_##B = _mm256_permute2x128_si256(x_##B, x_##B2, 0x20); \
t_##B2 = _mm256_permute2x128_si256(x_##B, x_##B2, 0x31); \
t_##C = _mm256_permute2x128_si256(x_##C, x_##C2, 0x20); \
t_##C2 = _mm256_permute2x128_si256(x_##C, x_##C2, 0x31); \
t_##D = _mm256_permute2x128_si256(x_##D, x_##D2, 0x20); \
t_##D2 = _mm256_permute2x128_si256(x_##D, x_##D2, 0x31); \
_mm256_storeu_si256((__m256i *)(c + 0), t_##A); \
_mm256_storeu_si256((__m256i *)(c + 64), t_##B); \
_mm256_storeu_si256((__m256i *)(c + 128), t_##C); \
_mm256_storeu_si256((__m256i *)(c + 192), t_##D); \
_mm256_storeu_si256((__m256i *)(c + 256), t_##A2); \
_mm256_storeu_si256((__m256i *)(c + 320), t_##B2); \
_mm256_storeu_si256((__m256i *)(c + 384), t_##C2); \
_mm256_storeu_si256((__m256i *)(c + 448), t_##D2); \
#define ONEOCTO(A, B, C, D, A2, B2, C2, D2, c) \
{ \
ONEQUAD_UNPCK(A, B, C, D); \
ONEQUAD_UNPCK(A2, B2, C2, D2); \
t[A] = _mm256_permute2x128_si256(x[A], x[A2], 0x20); \
t[A2] = _mm256_permute2x128_si256(x[A], x[A2], 0x31); \
t[B] = _mm256_permute2x128_si256(x[B], x[B2], 0x20); \
t[B2] = _mm256_permute2x128_si256(x[B], x[B2], 0x31); \
t[C] = _mm256_permute2x128_si256(x[C], x[C2], 0x20); \
t[C2] = _mm256_permute2x128_si256(x[C], x[C2], 0x31); \
t[D] = _mm256_permute2x128_si256(x[D], x[D2], 0x20); \
t[D2] = _mm256_permute2x128_si256(x[D], x[D2], 0x31); \
_mm256_storeu_si256((__m256i*)(c + 0), t[A]); \
_mm256_storeu_si256((__m256i*)(c + 64), t[B]); \
_mm256_storeu_si256((__m256i*)(c + 128), t[C]); \
_mm256_storeu_si256((__m256i*)(c + 192), t[D]); \
_mm256_storeu_si256((__m256i*)(c + 256), t[A2]); \
_mm256_storeu_si256((__m256i*)(c + 320), t[B2]); \
_mm256_storeu_si256((__m256i*)(c + 384), t[C2]); \
_mm256_storeu_si256((__m256i*)(c + 448), t[D2]); \
}
static inline uint64_t
_cha_8block(uint32_t* state, uint8_t* begin, uint8_t* end) {
if (end - begin < 512)
return 0;
uint8_t* c = begin;
/* constant for shuffling bytes (replacing multiple-of-8 rotates) */
__m256i rot16 = _mm256_set_epi8(
13, 12, 15, 14, 9, 8, 11, 10, 5, 4, 7, 6, 1, 0, 3, 2, 13, 12, 15, 14, 9,
8, 11, 10, 5, 4, 7, 6, 1, 0, 3, 2
);
__m256i rot8 = _mm256_set_epi8(
14, 13, 12, 15, 10, 9, 8, 11, 6, 5, 4, 7, 2, 1, 0, 3, 14, 13, 12, 15, 10,
9, 8, 11, 6, 5, 4, 7, 2, 1, 0, 3
);
/* the naive way seems as fast (if not a bit faster) than the vector way */
__m256i x[16], orig[16], t[16];
for (int i = 0; i < 16; ++i)
orig[i] = _mm256_set1_epi32(state[i]);
const __m256i addv12 = _mm256_set_epi64x(3, 2, 1, 0);
const __m256i addv13 = _mm256_set_epi64x(7, 6, 5, 4);
while (end - c >= 512) {
for (int i = 0; i < 16; ++i)
x[i] = orig[i];
// Calculate the eight parallel counters on x_12 and x_13
uint64_t* counter = (uint64_t*)(x + 12);
t[13] = _mm256_broadcastq_epi64(_mm_cvtsi64_si128(*counter));
t[12] = _mm256_add_epi64(addv12, t[13]);
t[13] = _mm256_add_epi64(addv13, t[13]);
x[12] = _mm256_unpacklo_epi32(t[12], t[13]);
x[13] = _mm256_unpackhi_epi32(t[12], t[13]);
t[12] = _mm256_unpacklo_epi32(x[12], x[13]);
t[13] = _mm256_unpackhi_epi32(x[12], x[13]);
/* required because unpack* are intra-lane */
const __m256i permute = _mm256_set_epi32(7, 6, 3, 2, 5, 4, 1, 0);
x[12] = _mm256_permutevar8x32_epi32(t[12], permute);
x[13] = _mm256_permutevar8x32_epi32(t[13], permute);
orig[12] = x[12];
orig[13] = x[13];
for (int i = 0; i < 10; ++i) {
VEC8_ROUND(0, 4, 8, 12, 1, 5, 9, 13, 2, 6, 10, 14, 3, 7, 11, 15);
VEC8_ROUND(0, 5, 10, 15, 1, 6, 11, 12, 2, 7, 8, 13, 3, 4, 9, 14);
}
ONEOCTO(0, 1, 2, 3, 4, 5, 6, 7, c);
ONEOCTO(8, 9, 10, 11, 12, 13, 14, 15, c + 32);
*counter += 8;
bytes -= 512;
c += 512;
}
return c - begin;
}
#undef ONEQUAD