Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

fix: MSVC SIMD compiling #3149

Open
wants to merge 3 commits into
base: master
Choose a base branch
from
Open

fix: MSVC SIMD compiling #3149

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
39f307a
fix: MSVC needs `AVX2` arch
EduMenges Jan 3, 2025
3e29dbb
fix: MNNSSE needs `AVX` arch in MSVC
EduMenges Jan 3, 2025
bbf755f
fix: Unified SIMD dereference accross compilers
EduMenges Jan 3, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
3 changes: 2 additions & 1 deletion source/backend/cpu/x86_x64/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,8 +86,9 @@ if(CMAKE_SYSTEM_PROCESSOR MATCHES "(x86_64)|(X86_64)|(x64)|(X64)|(amd64)|(AMD64)
target_compile_options(MNNAVXFMA PRIVATE -DMNN_USE_SSE)
endif()
if(MSVC)
target_compile_options(MNNSSE PRIVATE /arch:AVX)
if (MNN_AVX2)
target_compile_options(MNNAVX PRIVATE /arch:AVX)
target_compile_options(MNNAVX PRIVATE /arch:AVX2)
target_compile_options(MNNAVXFMA PRIVATE /arch:AVX2)
endif()
else()
Expand Down
10 changes: 3 additions & 7 deletions source/backend/cpu/x86_x64/avx/GemmInt8.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1308,13 +1308,9 @@ void _AVX_MNNFloat2Int8(const float* src, int8_t* dst, size_t sizeQuad, const fl
d0 = _mm256_add_epi32(d0, offset);
d0 = _mm256_packs_epi32(d0, _mm256_setzero_si256());
d0 = _mm256_permute4x64_epi64(d0, 0xD8);
#if defined(_MSC_VER)
__m256i x = static_cast<__m256i>(_mm256_packus_epi16(d0, _mm256_setzero_si256()));
*((int64_t*)dst + i) = x.m256i_i64[0];
#else
__v4di x = static_cast<__v4di>(_mm256_packus_epi16(d0, _mm256_setzero_si256()));
*((int64_t*)dst + i) = x[0];
#endif

__m256i x = _mm256_packus_epi16(d0, _mm256_setzero_si256());
*((int64_t*)dst + i) = _mm256_extract_epi64(x, 0);
}
}

Expand Down
Loading