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

Merge pull request #53 from Spudz76/ocean-master-revertCNGPUinterleave

Browse source 
Revert cn/gpu interleave (fixes ARM / Apple M1)
This commit is contained in:
MoneroOcean 2021-05-17 21:49:05 +00:00 • committed by GitHub
commit fef0e217c1
No known key found for this signature in database
GPG key ID: 4AEE18F83AFDEB23
2 changed files with 4 additions and 4 deletions
Download patch file Download diff file Expand all files Collapse all files

4
src/crypto/cn/CnHash.cpp
View file

@ -310,8 +310,8 @@ xmrig::CnHash::CnHash()
# endif # endif
# ifdef XMRIG_ALGO_CN_GPU # ifdef XMRIG_ALGO_CN_GPU
m_map[Algorithm::CN_GPU][AV_SINGLE][Assembly::NONE] = cryptonight_single_hash_gpu<Algorithm::CN_GPU, false, 0>; m_map[Algorithm::CN_GPU][AV_SINGLE][Assembly::NONE] = cryptonight_single_hash_gpu<Algorithm::CN_GPU, false>;
m_map[Algorithm::CN_GPU][AV_SINGLE_SOFT][Assembly::NONE] = cryptonight_single_hash_gpu<Algorithm::CN_GPU, true, 0>; m_map[Algorithm::CN_GPU][AV_SINGLE_SOFT][Assembly::NONE] = cryptonight_single_hash_gpu<Algorithm::CN_GPU, true>;
# endif # endif
# ifdef XMRIG_FEATURE_ASM # ifdef XMRIG_FEATURE_ASM

4
src/crypto/cn/CryptoNight_x86.h
View file

@ -830,7 +830,7 @@ void cn_explode_scratchpad_gpu(const uint8_t *input, uint8_t *output)
} }
template<Algorithm::Id ALGO, bool SOFT_AES, int interleave> template<Algorithm::Id ALGO, bool SOFT_AES>
inline void cryptonight_single_hash_gpu(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx, uint64_t) inline void cryptonight_single_hash_gpu(const uint8_t *__restrict__ input, size_t size, uint8_t *__restrict__ output, cryptonight_ctx **__restrict__ ctx, uint64_t)
{ {
constexpr CnAlgo<ALGO> props; constexpr CnAlgo<ALGO> props;
@ -850,7 +850,7 @@ inline void cryptonight_single_hash_gpu(const uint8_t *__restrict__ input, size_
cn_gpu_inner_ssse3<props.iterations(), props.mask()>(ctx[0]->state, ctx[0]->memory); cn_gpu_inner_ssse3<props.iterations(), props.mask()>(ctx[0]->state, ctx[0]->memory);
} }
cn_implode_scratchpad<ALGO, SOFT_AES, interleave>(reinterpret_cast<const __m128i *>(ctx[0]->memory), reinterpret_cast<__m128i *>(ctx[0]->state)); cn_implode_scratchpad<ALGO, SOFT_AES, 0>(reinterpret_cast<const __m128i *>(ctx[0]->memory), reinterpret_cast<__m128i *>(ctx[0]->state));
keccakf(reinterpret_cast<uint64_t*>(ctx[0]->state), 24); keccakf(reinterpret_cast<uint64_t*>(ctx[0]->state), 24);
memcpy(output, ctx[0]->state, 32); memcpy(output, ctx[0]->state, 32);
} }