[mxfp8 moe training] update readme with kernel microbenchmarks for dsv3

[danielvegamyhre]

Stacked PRs:

• ->[mxfp8 moe training] update readme with kernel microbenchmarks for dsv3 #3521
• [mxfp8 moe training] test floor vs rceil in 3d quantization benchmarks #3507
• [mxfp8 moe training] update benchmark plots to use new kernel #3506
• [mxfp8 moe training] integrate new cuda kernel for blocked layout for groups along K #3505
• [mxfp8 moe training] add CUDA kernel for per-group conversion of scale factors to blocked layout #3504

---

[mxfp8 moe training] update readme with kernel microbenchmarks for dsv3

Add mxfp8 kernel microbenchmarks for DeepSeekV3 shapes, grouped by forward/dgrad/wgrad. With environment details and commands to reproduce.

As you can see, the last weak point that meaningfully affects performance for DSV3 is the per grouped blocked layout kernel for groups along M (used in forward). We should be able to easily make a modified version of the new CUDA kernel added in the stack that operates on groups along K to instead operate for groups along M and resolve this issue.

Preview:

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/ao/3521

• 📄 Preview Python docs built from this PR

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

❌ 5 New Failures

As of commit 98e81e0 with merge base 7035fb7 ():

NEW FAILURES - The following jobs have failed:

• Run Regression Tests / test (CUDA 2.6, linux.g5.12xlarge.nvidia.gpu, torch==2.6.0, cuda, 12.6) / linux-job (gh)
  RuntimeError: Command docker exec -t f551cdf480a87f8534ef373813f94bb2e107914fa8449f6b598aa50f843796a7 /exec failed with exit code 2
• Run Regression Tests / test (CUDA 2.7, linux.g5.12xlarge.nvidia.gpu, torch==2.7.1, cuda, 12.6) / linux-job (gh)
  RuntimeError: Command docker exec -t 778ec38a4055a2f1e763f42740033d5fbc19a316250c5414efa3c2d96275d7e4 /exec failed with exit code 2
• Run Regression Tests / test (CUDA 2.8, linux.g5.12xlarge.nvidia.gpu, torch==2.8.0, cuda, 12.6) / linux-job (gh)
  RuntimeError: Command docker exec -t 83bd04a28956e694f3d156aaede09a5b90bf07c5b73ef41b8afd5bd2b904b988 /exec failed with exit code 2
• Run Regression Tests / test (CUDA 2.9, linux.g5.12xlarge.nvidia.gpu, torch==2.9.1, cuda, 12.6) / linux-job (gh)
  RuntimeError: Command docker exec -t 972ce0d5467f2c5d73c1ab44ac6c8b79d1b6865bfba0d726b5725feff5e1e381 /exec failed with exit code 2
• Run Regression Tests / test-nightly (CUDA Nightly, linux.g5.12xlarge.nvidia.gpu, --pre torch --index-url https://downloa... / linux-job (gh)
  RuntimeError: Command docker exec -t bccab297b96bcc33ded0e468f4f1646e3b7f3181228e20b617c0305bc76b57b8 /exec failed with exit code 2

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