mx: make CUDA kernel for dim1 cast in mxfp8_cublas recipe

[vkuzo]

Summary:

As titled, this is the fastest option right now so should be default.

Test Plan:

pytest test/prototype/mx_formats/ -s -x

Reviewers:

Subscribers:

Tasks:

Tags:

[vkuzo]

Stack from ghstack (oldest at bottom):

• -> mx: make CUDA kernel for dim1 cast in mxfp8_cublas recipe #2661

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/2661

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

⏳ 13 Pending, 1 Unrelated Failure

As of commit aa7c790 with merge base c993d64 ():

FLAKY - The following job failed but was likely due to flakiness present on trunk:

• Run 1xH100 Tests / test (H100, linux.aws.h100, --pre torch torchvision torchaudio --index-url https://download.pytor... / linux-job (gh) (matched linux rule in flaky-rules.json)
  The process '/usr/bin/git' failed with exit code 128

This comment was automatically generated by Dr. CI and updates every 15 minutes.