Skip to content

Improve add loop #193

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

Draft
antonwolfy wants to merge 7 commits into
base: main
Choose a base branch
from
Draft

Improve add loop #193

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
7 commits
Select commit Hold shift + click to select a range
4280d9a
Implement VEX Encoding for add function to have AVX2 add loop similar…
antonwolfy Mar 26, 2026
a725d7a
Use MKL VML LinearFrac to fix regression observing in go_fast benchmark
antonwolfy Mar 26, 2026
5c4d377
Add reduction branch back to add function
antonwolfy Mar 26, 2026
81f6c4b
Undo the temporary changes in conda-recipe/meta.yaml
antonwolfy Apr 14, 2026
5ddffbc
Remove debug files
antonwolfy Apr 14, 2026
0205909
Revert logic to match Intel compiler in CMakeLists.txt
antonwolfy Apr 14, 2026
7038e56
Remove rebase leftover
antonwolfy Apr 14, 2026
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
6 changes: 5 additions & 1 deletion CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,14 +109,18 @@ set_target_properties(${_trgt} PROPERTIES
CMAKE_POSITION_INDEPENDENT_CODE ON
C_STANDARD 99
)
target_include_directories(${_trgt} PUBLIC mkl_umath/src/ ${Python_NumPy_INCLUDE_DIRS} ${Python_INCLUDE_DIRS})
target_include_directories(${_trgt} PUBLIC mkl_umath/src/ mkl_umath/src/npyv/ ${Python_NumPy_INCLUDE_DIRS} ${Python_INCLUDE_DIRS})
target_link_libraries(${_trgt} PUBLIC MKL::MKL ${Python_LIBRARIES})
target_link_options(${_trgt} PUBLIC ${_linker_options})
target_compile_options(${_trgt} PUBLIC -fveclib=SVML)
target_compile_options(${_trgt} PUBLIC -fvectorize)
if(OPTIMIZATION_REPORT)
target_compile_options(${_trgt} PRIVATE -qopt-report=3)
endif()
# Enable AVX2 for npyv SIMD acceleration
# -mvzeroupper: Prefer VEX encoding even for scalar ops
# -mtune=native: Optimize for build machine (can change to -mtune=haswell for generic AVX2)
target_compile_options(${_trgt} PUBLIC -mavx2 -mfma -mvzeroupper)
install(TARGETS ${_trgt}
LIBRARY DESTINATION mkl_umath
ARCHIVE DESTINATION mkl_umath
Expand Down
Loading
Loading