GH-48160: [C++][Gandiva] Pass CPU attributes to LLVM

[lriggs]

Rationale for this change

The CPU attributes are not passed to the LLVM layer, which means potential optimizations could be missed leading to inefficient code. This feature was lost as part of the refactoring in 83cba25

I also discovered a bug with decimal alignments that was exposed by this change and was only reproducible in our test environment.

What changes are included in this PR?

Pass the CPU attributes to the LLVM code generation, and a unit test.
Fix the 16 bit vs 8 bit decimal alignment problem. This was causing a crash sometimes on certain architectures with certain queries. Added a unit test.

Are these changes tested?

Yes.

Are there any user-facing changes?

No.

• GitHub Issue: [C++][Gandiva] CPU attributes are not passed to the LLVM layer during code generation #48160

[github-actions]

⚠️ GitHub issue #48160 has been automatically assigned in GitHub to PR creator.

[raulcd]

@github-actions crossbow submit -g cpp

[github-actions]

Revision: 704804b

Submitted crossbow builds: ursacomputing/crossbow @ actions-2b90f7a897

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-cuda-cpp-ubuntu-22.04-cuda-11.7.1
test-debian-12-cpp-amd64
test-debian-12-cpp-i386
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

[kou]

+1

[conbench-apache-arrow]

After merging your PR, Conbench analyzed the 3 benchmarking runs that have been run so far on merge-commit 7bc47c0.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details.