GH-49449: [C++] Backport xsimd neon fix

[AntoinePrv]

Rationale for this change

Performance in hot code.
There is a bug in xsimd where the rshift/lshift for Neon are implemented with a scalar loop instead of the appropriate SIMD intrinsics. This code path is core to the unpack routine in Parquet reads.
xtensor-stack/xsimd#1266

What changes are included in this PR?

A bug backport.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

• GitHub Issue: [C++] Backport Neon xsimd fix #49449

[github-actions]

⚠️ GitHub issue #49449 has been automatically assigned in GitHub to PR creator.

[AntoinePrv]

@pitrou

[pitrou]

Can you explain what this is fixing in the PR description?

[pitrou]

Also, were you able to run some benchmarks?

[AntoinePrv]

Also, were you able to run some benchmarks?

Not yet, I will do it overnight (I run some extensive benchmarks in a fork). I was however able to confirm from the assembly that this is the correct instructions we aimed to use.

[AntoinePrv]

@pitrou I've run the benchmarks (though I have lost the old ones in the process) and it is much better.
It went from being on par with the scalar / old version to being much better.

[pitrou]

@github-actions crossbow submit -g cpp

[github-actions]

Revision: 18ce22f

Submitted crossbow builds: ursacomputing/crossbow @ actions-471834543b

Task	Status
example-cpp-minimal-build-static
example-cpp-minimal-build-static-system-dependency
example-cpp-tutorial
test-build-cpp-fuzz
test-conda-cpp
test-conda-cpp-valgrind
test-debian-13-cpp-amd64
test-debian-13-cpp-i386
test-debian-experimental-cpp-gcc-15
test-fedora-42-cpp
test-ubuntu-22.04-cpp
test-ubuntu-22.04-cpp-20
test-ubuntu-22.04-cpp-bundled
test-ubuntu-22.04-cpp-emscripten
test-ubuntu-22.04-cpp-no-threading
test-ubuntu-24.04-cpp
test-ubuntu-24.04-cpp-bundled-offline
test-ubuntu-24.04-cpp-gcc-13-bundled
test-ubuntu-24.04-cpp-gcc-14
test-ubuntu-24.04-cpp-minimal-with-formats
test-ubuntu-24.04-cpp-thread-sanitizer

[pitrou]

@ursabot please benchmark lang=C++

[pitrou]

+1, I trust you on the benchmarks given that our comment bot seem out of order (@rok). Will merge.

[rok]

@ursabot please benchmark lang=C++

[rok]

Benchmark runs are scheduled for commit 18ce22f. Watch https://buildkite.com/apache-arrow and https://conbench.arrow-dev.org for updates. A comment will be posted here when the runs are complete.

[rok]

I didn't realize we already cover neon via graviton machines, that's nice.

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit dbbf7cf.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 26 possible false positives for unstable benchmarks that are known to sometimes produce them.

[rok]

@pitrou benchmarks look ok.

[conbench-apache-arrow]

Thanks for your patience. Conbench analyzed the 3 benchmarking runs that have been run so far on PR commit 18ce22f.

There were 7 benchmark results indicating a performance regression:

• Pull Request Run on amd64-c6a-4xlarge-linux at 2026-03-09 16:32:46Z
  • BM_UnpackUint64 (C++) with params=DynamicUnaligned/8/256, source=cpp-micro, suite=arrow-bpacking-benchmark
  • BM_UnpackUint64 (C++) with params=DynamicAligned/8/256, source=cpp-micro, suite=arrow-bpacking-benchmark
• and 5 more (see the report linked below)

The full Conbench report has more details.