speed up nvte_multi_padding / nvte_multi_unpadding

[matthiasdiener]

Description

Please include a brief summary of the changes, relevant motivation and context.

Fixes https://github.com/ROCm/frameworks-internal/issues/16530

See https://github.com/ROCm/frameworks-internal/issues/16530#issuecomment-4502138388 for performance.

Type of change

• Documentation change (change only to the documentation, either a fix or a new content)
• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Infra/Build change
• Code refactoring

Changes

Please list the changes introduced in this PR:

• Change A
• Change B

Checklist:

• I have read and followed the contributing guidelines
• The functionality is complete
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[aris134]

LGTM!

[alextmagro]

What is the case where we hit non-aligned vectors? Isn't FP8/MXFP8 always padded to a multiple of 16 by default? Ideally we would template out the NT vs elementwise stores

[alextmagro]

Did you try #pragma unroll 2 here? If we have the registers available that might help performance.