fzorb/REDACTED-rig
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Reverted accidentaly reverted stuff

Browse source
This commit is contained in:
BenDr0id 2019-03-06 13:52:04 +01:00
parent ad93b74481
commit 75125f22d3
1 changed files with 7 additions and 2 deletions
Download patch file Download diff file Expand all files Collapse all files

9
src/crypto/CryptoNight_x86.h
View file

@ -195,9 +195,9 @@ static inline uint64_t __umul128(uint64_t multiplier, uint64_t multiplicand, uin
# define SHUFFLE_PHASE_1(l, idx, bx0, bx1, ax, reverse) \ # define SHUFFLE_PHASE_1(l, idx, bx0, bx1, ax, reverse) \
{ \ { \
const __m128i chunk1 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x10))); \ const __m128i chunk1 = _mm_load_si128((__m128i *)((l) + ((idx) ^ (reverse ? 0x30 : 0x10)))); \
const __m128i chunk2 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x20))); \ const __m128i chunk2 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x20))); \
const __m128i chunk3 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x30))); \ const __m128i chunk3 = _mm_load_si128((__m128i *)((l) + ((idx) ^ (reverse ? 0x10 : 0x30)))); \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x10)), _mm_add_epi64(chunk3, bx1)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x10)), _mm_add_epi64(chunk3, bx1)); \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x20)), _mm_add_epi64(chunk1, bx0)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x20)), _mm_add_epi64(chunk1, bx0)); \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x30)), _mm_add_epi64(chunk2, ax)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x30)), _mm_add_epi64(chunk2, ax)); \
@ -221,8 +221,13 @@ static inline uint64_t __umul128(uint64_t multiplier, uint64_t multiplicand, uin
const __m128i chunk3 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x30))); \ const __m128i chunk3 = _mm_load_si128((__m128i *)((l) + ((idx) ^ 0x30))); \
hi ^= ((uint64_t*)((l) + ((idx) ^ 0x20)))[0]; \ hi ^= ((uint64_t*)((l) + ((idx) ^ 0x20)))[0]; \
lo ^= ((uint64_t*)((l) + ((idx) ^ 0x20)))[1]; \ lo ^= ((uint64_t*)((l) + ((idx) ^ 0x20)))[1]; \
if (reverse) { \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x10)), _mm_add_epi64(chunk1, bx1)); \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x20)), _mm_add_epi64(chunk3, bx0)); \
} else { \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x10)), _mm_add_epi64(chunk3, bx1)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x10)), _mm_add_epi64(chunk3, bx1)); \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x20)), _mm_add_epi64(chunk1, bx0)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x20)), _mm_add_epi64(chunk1, bx0)); \
} \
_mm_store_si128((__m128i *)((l) + ((idx) ^ 0x30)), _mm_add_epi64(chunk2, ax)); \ _mm_store_si128((__m128i *)((l) + ((idx) ^ 0x30)), _mm_add_epi64(chunk2, ax)); \
} }