GH-44098: [C++] Add home made _mm256_set_m128i for compilers who are missing it

[zanmato1984]

Rationale for this change

AVX2 intrinsic _mm256_set_m128i is missing in GCC versions <8.0 - this is a GCC bug discussed in [1], causing certain CI build failed.

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=80582

What changes are included in this PR?

Check the GCC version and use a home made replacement if necessary.

Are these changes tested?

Manually tested.

Are there any user-facing changes?

None.

• GitHub Issue: [C++][Compute] _mm256_set_m128i() is unavailable on OpenSUSE 15.1 (GCC 7.5) #44098

[github-actions]

⚠️ GitHub issue #40098 has been automatically assigned in GitHub to PR creator.

[zanmato1984]

cc @kou @assignUser @jonkeane

[zanmato1984]

@github-actions crossbow submit test-r-rstudio-r-base-4.1-opensuse155

[github-actions]

Revision: 9134050

Submitted crossbow builds: ursacomputing/crossbow @ actions-0fe409cd3d

Task	Status
test-r-rstudio-r-base-4.1-opensuse155

[zanmato1984]

I try to find some common place to put simd specific code and found this one. However it seems this file is only adapting the static simd optimizations, thus the additions of the ARROW_HAVE_RUNTIME_* which are for dynamic simd optimizations.

W/o these || defined(ARROW_HAVE_RUNTIME_*, one won't see <immintrin.h> inclusion because there are no ARROW_HAVE_* macros defined. (This might be yet another issue of our current simd macro organization).

[zanmato1984]

Also cc @pitrou .

[github-actions]

⚠️ GitHub issue #44098 has been automatically assigned in GitHub to PR creator.

[pitrou]

@ursabot please benchmark lang=C++

[ursabot]

Benchmark runs are scheduled for commit 9134050. Watch https://buildkite.com/apache-arrow and https://conbench.ursa.dev for updates. A comment will be posted here when the runs are complete.

[conbench-apache-arrow]

Thanks for your patience. Conbench analyzed the 4 benchmarking runs that have been run so far on PR commit 9134050.

There were 15 benchmark results indicating a performance regression:

• Pull Request Run on arm64-t4g-2xlarge-linux at 2024-09-14 10:03:07Z
  • BitBlockCounterFourWordsSum (C++) with params=4096, source=cpp-micro, suite=arrow-bit-block-counter-benchmark
  • BenchmarkBitmapVisitBitsetAnd (C++) with params=131072/0, source=cpp-micro, suite=arrow-bit-util-benchmark
• and 13 more (see the report linked below)

The full Conbench report has more details.

[pitrou]

The so-called performance regressions are just flukes.

[pitrou]

@github-actions crossbow submit -g cpp -g r -g wheel

[github-actions]

Revision: 9134050

Submitted crossbow builds: ursacomputing/crossbow @ actions-8d654932f7

Task	Status
example-cpp-minimal-build-static
example-cpp-minimal-build-static-system-dependency
example-cpp-tutorial
python-sdist
r-binary-packages
test-alpine-linux-cpp
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-cuda-cpp
test-debian-12-cpp-amd64
test-debian-12-cpp-i386
test-fedora-39-cpp
test-r-arrow-backwards-compatibility
test-r-clang-sanitizer
test-r-depsource-bundled
test-r-depsource-system
test-r-dev-duckdb
test-r-devdocs
test-r-extra-packages
test-r-gcc-11
test-r-gcc-12
test-r-install-local
test-r-install-local-minsizerel
test-r-linux-as-cran
test-r-linux-rchk
test-r-linux-valgrind
test-r-macos-as-cran
test-r-minimal-build
test-r-offline-maximal
test-r-offline-minimal
test-r-rhub-debian-gcc-devel-lto-latest
test-r-rhub-debian-gcc-release-custom-ccache
test-r-rhub-ubuntu-release-latest
test-r-rocker-r-ver-latest
test-r-rstudio-r-base-4.1-opensuse155
test-r-rstudio-r-base-4.2-focal
test-r-ubuntu-22.04
test-r-versions
test-ubuntu-20.04-cpp
test-ubuntu-20.04-cpp-bundled
test-ubuntu-20.04-cpp-minimal-with-formats
test-ubuntu-20.04-cpp-thread-sanitizer
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-20
test-ubuntu-22.04-cpp-emscripten
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-gcc-13-bundled
test-ubuntu-24.04-cpp-gcc-14
test-ubuntu-r-sanitizer
wheel-macos-monterey-cp310-amd64
wheel-macos-monterey-cp310-arm64
wheel-macos-monterey-cp311-amd64
wheel-macos-monterey-cp311-arm64
wheel-macos-monterey-cp312-amd64
wheel-macos-monterey-cp312-arm64
wheel-macos-monterey-cp313-amd64
wheel-macos-monterey-cp313-arm64
wheel-macos-monterey-cp38-amd64
wheel-macos-monterey-cp38-arm64
wheel-macos-monterey-cp39-amd64
wheel-macos-monterey-cp39-arm64
wheel-manylinux-2-28-cp310-amd64
wheel-manylinux-2-28-cp310-arm64
wheel-manylinux-2-28-cp311-amd64
wheel-manylinux-2-28-cp311-arm64
wheel-manylinux-2-28-cp312-amd64
wheel-manylinux-2-28-cp312-arm64
wheel-manylinux-2-28-cp313-amd64
wheel-manylinux-2-28-cp313-arm64
wheel-manylinux-2-28-cp38-amd64
wheel-manylinux-2-28-cp38-arm64
wheel-manylinux-2-28-cp39-amd64
wheel-manylinux-2-28-cp39-arm64
wheel-manylinux-2014-cp310-amd64
wheel-manylinux-2014-cp310-arm64
wheel-manylinux-2014-cp311-amd64
wheel-manylinux-2014-cp311-arm64
wheel-manylinux-2014-cp312-amd64
wheel-manylinux-2014-cp312-arm64
wheel-manylinux-2014-cp313-amd64
wheel-manylinux-2014-cp313-arm64
wheel-manylinux-2014-cp38-amd64
wheel-manylinux-2014-cp38-arm64
wheel-manylinux-2014-cp39-amd64
wheel-manylinux-2014-cp39-arm64
wheel-windows-cp310-amd64
wheel-windows-cp311-amd64
wheel-windows-cp312-amd64
wheel-windows-cp313-amd64
wheel-windows-cp38-amd64
wheel-windows-cp39-amd64

[mapleFU]

Nice catch? Can we also try to try this to a gcc 7.5 and godbolt?

[pitrou]

+1, thank you @zanmato1984

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 87d6477.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 2 possible false positives for unstable benchmarks that are known to sometimes produce them.