FEX-Emu/xxHash
1
0
Fork 0
mirror of https://github.com/FEX-Emu/xxHash.git synced 2024-11-28 01:00:56 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

minor refactoring, for clarity

Browse Source
This commit is contained in:
Yann Collet 2019-07-17 17:53:10 -07:00
parent e0bfe4440f
commit 81410ce843
1 changed files with 18 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

28
xxh3.h
View File

@ -537,7 +537,7 @@ XXH3_scrambleAcc(void* XXH_RESTRICT acc, const void* XXH_RESTRICT key)
assert(((size_t)acc) & 31 == 0); assert(((size_t)acc) & 31 == 0);
{ XXH_ALIGN(32) __m256i* const xacc = (__m256i*) acc; { XXH_ALIGN(32) __m256i* const xacc = (__m256i*) acc;
const __m256i* const xkey = (const __m256i *) key; /* not really aligned, just for ptr arithmetic, and because _mm256_loadu_si256() requires this argument type */ const __m256i* const xkey = (const __m256i *) key; /* not really aligned, just for ptr arithmetic, and because _mm256_loadu_si256() requires this argument type */
const __m256i k1 = _mm256_set1_epi32((int)PRIME32_1); const __m256i prime32 = _mm256_set1_epi32((int)PRIME32_1);
size_t i; size_t i;
for (i=0; i < STRIPE_LEN/sizeof(__m256i); i++) { for (i=0; i < STRIPE_LEN/sizeof(__m256i); i++) {
@ -546,12 +546,12 @@ XXH3_scrambleAcc(void* XXH_RESTRICT acc, const void* XXH_RESTRICT key)
data = _mm256_xor_si256(data, shifted); data = _mm256_xor_si256(data, shifted);
{ __m256i const k = _mm256_loadu_si256 (xkey+i); { __m256i const k = _mm256_loadu_si256 (xkey+i);
__m256i const dk = _mm256_xor_si256 (data, k); /* U32 dk[4] = {d0+k0, d1+k1, d2+k2, d3+k3} */ __m256i const dk = _mm256_xor_si256 (data, k);
__m256i const dk1 = _mm256_mul_epu32 (dk, k1); __m256i const dk1 = _mm256_mul_epu32 (dk, prime32);
__m256i const d2 = _mm256_shuffle_epi32 (dk, 0x31); __m256i const d2 = _mm256_shuffle_epi32 (dk, 0x31);
__m256i const dk2 = _mm256_mul_epu32 (d2, k1); __m256i const dk2 = _mm256_mul_epu32 (d2, prime32);
__m256i const dk2h= _mm256_slli_epi64 (dk2, 32); __m256i const dk2h= _mm256_slli_epi64 (dk2, 32);
xacc[i] = _mm256_add_epi64(dk1, dk2h); xacc[i] = _mm256_add_epi64(dk1, dk2h);
@ -562,7 +562,7 @@ XXH3_scrambleAcc(void* XXH_RESTRICT acc, const void* XXH_RESTRICT key)
{ XXH_ALIGN(16) __m128i* const xacc = (__m128i*) acc; { XXH_ALIGN(16) __m128i* const xacc = (__m128i*) acc;
const __m128i* const xkey = (const __m128i *) key; /* not really aligned, just for ptr arithmetic */ const __m128i* const xkey = (const __m128i *) key; /* not really aligned, just for ptr arithmetic */
const __m128i k1 = _mm_set1_epi32((int)PRIME32_1); const __m128i prime32 = _mm_set1_epi32((int)PRIME32_1);
size_t i; size_t i;
for (i=0; i < STRIPE_LEN/sizeof(__m128i); i++) { for (i=0; i < STRIPE_LEN/sizeof(__m128i); i++) {
@ -573,10 +573,10 @@ XXH3_scrambleAcc(void* XXH_RESTRICT acc, const void* XXH_RESTRICT key)
{ __m128i const k = _mm_loadu_si128 (xkey+i); { __m128i const k = _mm_loadu_si128 (xkey+i);
__m128i const dk = _mm_xor_si128 (data,k); __m128i const dk = _mm_xor_si128 (data,k);
__m128i const dk1 = _mm_mul_epu32 (dk,k1); __m128i const dk1 = _mm_mul_epu32 (dk, prime32);
__m128i const d2 = _mm_shuffle_epi32 (dk, 0x31); __m128i const d2 = _mm_shuffle_epi32 (dk, 0x31);
__m128i const dk2 = _mm_mul_epu32 (d2,k1); __m128i const dk2 = _mm_mul_epu32 (d2, prime32);
__m128i const dk2h= _mm_slli_epi64(dk2, 32); __m128i const dk2h= _mm_slli_epi64(dk2, 32);
xacc[i] = _mm_add_epi64(dk1, dk2h); xacc[i] = _mm_add_epi64(dk1, dk2h);
@ -771,7 +771,8 @@ XXH3_hashLong_64b_defaultSecret(const void* XXH_RESTRICT data, size_t len)
} }
XXH_NO_INLINE XXH64_hash_t /* It's important for performance that XXH3_hashLong is not inlined. Not sure why (uop cache maybe ?), but difference is large and easily measurable */ XXH_NO_INLINE XXH64_hash_t /* It's important for performance that XXH3_hashLong is not inlined. Not sure why (uop cache maybe ?), but difference is large and easily measurable */
XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT data, size_t len, const void* XXH_RESTRICT secret, size_t secretSize) XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT data, size_t len,
const void* XXH_RESTRICT secret, size_t secretSize)
{ {
return XXH3_hashLong_internal(data, len, secret, secretSize); return XXH3_hashLong_internal(data, len, secret, secretSize);
} }
@ -818,7 +819,8 @@ XXH3_hashLong_64b_withSeed(const void* data, size_t len, XXH64_hash_t seed)
} }
XXH_FORCE_INLINE U64 XXH3_mix16B(const void* XXH_RESTRICT data, const void* XXH_RESTRICT key, U64 seed64) XXH_FORCE_INLINE U64 XXH3_mix16B(const void* XXH_RESTRICT data,
const void* XXH_RESTRICT key, U64 seed64)
{ {
const U64* const key64 = (const U64*)key; const U64* const key64 = (const U64*)key;
U64 const ll1 = XXH_readLE64(data); U64 const ll1 = XXH_readLE64(data);
@ -830,7 +832,9 @@ XXH_FORCE_INLINE U64 XXH3_mix16B(const void* XXH_RESTRICT data, const void* XXH_
XXH_FORCE_INLINE XXH64_hash_t XXH_FORCE_INLINE XXH64_hash_t
XXH3_len_17to128_64b(const void* XXH_RESTRICT data, size_t len, const void* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) XXH3_len_17to128_64b(const void* XXH_RESTRICT data, size_t len,
const void* XXH_RESTRICT secret, size_t secretSize,
XXH64_hash_t seed)
{ {
const BYTE* const p = (const BYTE*)data; const BYTE* const p = (const BYTE*)data;
const char* const key = (const char*)secret; const char* const key = (const char*)secret;
@ -859,7 +863,9 @@ XXH3_len_17to128_64b(const void* XXH_RESTRICT data, size_t len, const void* XXH_
} }
XXH_NO_INLINE XXH64_hash_t XXH_NO_INLINE XXH64_hash_t
XXH3_len_129to240_64b(const void* XXH_RESTRICT data, size_t len, const void* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) XXH3_len_129to240_64b(const void* XXH_RESTRICT data, size_t len,
const void* XXH_RESTRICT secret, size_t secretSize,
XXH64_hash_t seed)
{ {
const BYTE* const p = (const BYTE*)data; const BYTE* const p = (const BYTE*)data;
const char* const key = (const char*)secret; const char* const key = (const char*)secret;