Don't save fp8 q/k/v/out tensors when using bf16 bprop

[guyueh1]

Description

When using FP8_DPA=1 NVTE_FP8_DPA_BWD=0, the backprop uses BF16 q/k/v/out tensors and the fp8 q/k/v/o are not used. So we should avoid saving them for backprop, which reduces the peak memory footprint.

Fixes # (issue)

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 refractor

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

[ksivaman]

/te-ci pytorch