Skip to content

GH-45331: [C++] Use xsimd for CPU feature detection #49940

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

Merged
pitrou merged 6 commits into from
Jun 10, 2026
+181 −408
Merged

GH-45331: [C++] Use xsimd for CPU feature detection #49940

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
6 commits
Select commit Hold shift + click to select a range
5454426
Use xsimf in CpuInfo and remove Pimpl
AntoinePrv May 6, 2026
6ca9c33
Fix headers
AntoinePrv May 7, 2026
bdc923a
Fix convertion
AntoinePrv May 7, 2026
557d65b
Fix comments
AntoinePrv Jun 3, 2026
f5cab60
Make xsimd mandatory
AntoinePrv Jun 5, 2026
feda2a5
Handle emscripten
AntoinePrv Jun 5, 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
34 changes: 13 additions & 21 deletions cpp/cmake_modules/ThirdpartyToolchain.cmake
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -2844,28 +2844,20 @@ macro(build_xsimd)
set(XSIMD_VENDORED TRUE)
endmacro()

if((NOT ARROW_SIMD_LEVEL STREQUAL "NONE") OR (NOT ARROW_RUNTIME_SIMD_LEVEL STREQUAL "NONE"
))
set(ARROW_USE_XSIMD TRUE)
# Xsimd is mandatory as its CPU feature detection is the basis for Arrow CpuInfo
resolve_dependency(xsimd
FORCE_ANY_NEWER_VERSION
TRUE
IS_RUNTIME_DEPENDENCY
FALSE
REQUIRED_VERSION
"14.2.0")

if(xsimd_SOURCE STREQUAL "BUNDLED")
set(ARROW_XSIMD arrow::xsimd)
else()
set(ARROW_USE_XSIMD FALSE)
endif()

if(ARROW_USE_XSIMD)
resolve_dependency(xsimd
FORCE_ANY_NEWER_VERSION
TRUE
IS_RUNTIME_DEPENDENCY
FALSE
REQUIRED_VERSION
"14.2.0")

if(xsimd_SOURCE STREQUAL "BUNDLED")
set(ARROW_XSIMD arrow::xsimd)
else()
message(STATUS "xsimd found. Headers: ${xsimd_INCLUDE_DIRS}")
set(ARROW_XSIMD xsimd)
endif()
message(STATUS "xsimd found. Headers: ${xsimd_INCLUDE_DIRS}")
set(ARROW_XSIMD xsimd)
endif()

macro(build_zlib)
Expand Down
30 changes: 11 additions & 19 deletions cpp/src/arrow/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -601,6 +601,9 @@ arrow_add_object_library(ARROW_UTIL ${ARROW_UTIL_SRCS})
foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
target_compile_definitions(${ARROW_UTIL_TARGET} PRIVATE URI_STATIC_BUILD)
endforeach()
foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()

if(ARROW_USE_BOOST)
foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
Expand All @@ -617,11 +620,6 @@ if(ARROW_USE_OPENSSL)
target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE ${ARROW_OPENSSL_LIBS})
endforeach()
endif()
if(ARROW_USE_XSIMD)
foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()
endif()
if(ARROW_WITH_BROTLI)
foreach(ARROW_UTIL_TARGET ${ARROW_UTIL_TARGETS})
target_link_libraries(${ARROW_UTIL_TARGET} PRIVATE ${ARROW_BROTLI_LIBS})
Expand Down Expand Up @@ -734,11 +732,9 @@ if(ARROW_CSV)
csv/parser.cc
csv/reader.cc
csv/writer.cc)
if(ARROW_USE_XSIMD)
foreach(ARROW_CSV_TARGET ${ARROW_CSV_TARGETS})
target_link_libraries(${ARROW_CSV_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()
endif()
foreach(ARROW_CSV_TARGET ${ARROW_CSV_TARGETS})
target_link_libraries(${ARROW_CSV_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()

list(APPEND ARROW_TESTING_SRCS csv/test_common.cc)
else()
Expand Down Expand Up @@ -855,15 +851,13 @@ if(ARROW_COMPUTE)
list(APPEND ARROW_COMPUTE_SHARED_INSTALL_INTERFACE_LIBS Arrow::arrow_shared)
list(APPEND ARROW_COMPUTE_STATIC_LINK_LIBS arrow_static)
list(APPEND ARROW_COMPUTE_SHARED_LINK_LIBS arrow_shared)
list(APPEND ARROW_COMPUTE_STATIC_LINK_LIBS ${ARROW_XSIMD})
list(APPEND ARROW_COMPUTE_SHARED_PRIVATE_LINK_LIBS ${ARROW_XSIMD})
Comment on lines +854 to +855

@pitrou pitrou Jun 9, 2026

Copy link
Copy Markdown
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't see xsimd in use anywhere in arrow/compute, do you remember why this was required @raulcd ?

Edit: perhaps because of src/arrow/util/utf8_internal.h? It's a pity this indirect dependency isn't picked up automatically.


if(ARROW_USE_BOOST)
list(APPEND ARROW_COMPUTE_STATIC_LINK_LIBS Boost::headers)
list(APPEND ARROW_COMPUTE_SHARED_PRIVATE_LINK_LIBS Boost::headers)
endif()
if(ARROW_USE_XSIMD)
list(APPEND ARROW_COMPUTE_STATIC_LINK_LIBS ${ARROW_XSIMD})
list(APPEND ARROW_COMPUTE_SHARED_PRIVATE_LINK_LIBS ${ARROW_XSIMD})
endif()
if(ARROW_WITH_OPENTELEMETRY)
list(APPEND ARROW_COMPUTE_STATIC_LINK_LIBS ${ARROW_OPENTELEMETRY_LIBS})
list(APPEND ARROW_COMPUTE_SHARED_PRIVATE_LINK_LIBS ${ARROW_OPENTELEMETRY_LIBS})
Expand Down Expand Up @@ -910,11 +904,9 @@ endif()

arrow_add_object_library(ARROW_COMPUTE_CORE ${ARROW_COMPUTE_SRCS})

if(ARROW_USE_XSIMD)
foreach(ARROW_COMPUTE_CORE_TARGET ${ARROW_COMPUTE_CORE_TARGETS})
target_link_libraries(${ARROW_COMPUTE_CORE_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()
endif()
foreach(ARROW_COMPUTE_CORE_TARGET ${ARROW_COMPUTE_CORE_TARGETS})
target_link_libraries(${ARROW_COMPUTE_CORE_TARGET} PRIVATE ${ARROW_XSIMD})
endforeach()
if(ARROW_WITH_OPENTELEMETRY)
foreach(ARROW_COMPUTE_CORE_TARGET ${ARROW_COMPUTE_CORE_TARGETS})
target_link_libraries(${ARROW_COMPUTE_CORE_TARGET}
Expand Down
8 changes: 2 additions & 6 deletions cpp/src/arrow/util/CMakeLists.txt
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -39,9 +39,7 @@ else()
endif()

set(ARROW_UTILITY_TEST_LINK_LIBS Boost::headers)
if(ARROW_USE_XSIMD)
list(APPEND ARROW_UTILITY_TEST_LINK_LIBS ${ARROW_XSIMD})
endif()
list(APPEND ARROW_UTILITY_TEST_LINK_LIBS ${ARROW_XSIMD})
if(ARROW_WITH_OPENTELEMETRY)
list(APPEND ARROW_UTILITY_TEST_LINK_LIBS ${ARROW_OPENTELEMETRY_LIBS})
endif()
Expand Down Expand Up @@ -135,9 +133,7 @@ add_arrow_benchmark(tdigest_benchmark)
add_arrow_benchmark(thread_pool_benchmark)
add_arrow_benchmark(trie_benchmark)
set(ARROW_BENCHMARK_UTF8_UTIL_LINK_LIBS)
if(ARROW_USE_XSIMD)
list(APPEND ARROW_BENCHMARK_UTF8_UTIL_LINK_LIBS ${ARROW_XSIMD})
endif()
list(APPEND ARROW_BENCHMARK_UTF8_UTIL_LINK_LIBS ${ARROW_XSIMD})
add_arrow_benchmark(utf8_util_benchmark EXTRA_LINK_LIBS
${ARROW_BENCHMARK_UTF8_UTIL_LINK_LIBS})
add_arrow_benchmark(value_parsing_benchmark)
Loading
Loading