Fix ARM half conversion in RenderOutputWriter

[gripeyes]

Compatibility:

patch

Issues/Tickets:

Render AOVs from Hou on Silicon macOS contained NaN without this fix, this fix got it working.

Release notes comment:

Fixed scalar half-float conversion in RenderOutputWriter on Apple Silicon.

Comments for the reviewer:

The previous ARM/NEON implementation used vector load/store intrinsics with scalar addresses. This change converts through scalar __fp16 values and uses std::memcpy to preserve the exact 16-bit representation.

This is one part of an earlier Apple Silicon native AOV investigation:
OpenMoonRay/hdMoonray@c7bbd30...4234f27

The separate hdMoonray render-buffer allocation change from that investigation is intentionally excluded from this PR. This PR retains only the independent low-level Apple Silicon half-conversion correction.

Look or scene setup change:

No scene setup changes are required. On Apple Silicon, outputs previously affected by incorrect half-float conversion may now contain the intended values.

Special notes for production:

The changed code path is limited to ARM/NEON builds. The existing non-ARM conversion path is unchanged.

Attention/Reviewers:

AI Assisted Development:

Assisted-by: OpenAI Codex / GPT-5.6

Checklist:

• Documentation has been updated.
• Includes new unit tests.
• Includes new RATS tests.

No documentation, unit-test, or RATS changes are included in this narrowly scoped correction.

[gripeyes]

Also mentioning this one since they're both related:
OpenMoonRay/scene_rdl2#25

Please let me know if this makes sense and if PRs like this are expected/ideal.