MX4 benchmark

[spcyppt]

Summary: MX4 benchmark to compare performance of MX4 and FP8 quantize/dequantize

Usage:

buck2 run @//mode/opt //deeplearning/fbgemm/fbgemm_gpu:quantize_ops_benchmark -- bench-mx4

Arguments

    --enable_trace_profile <bool> # enables kineto kernel profile, default=False
    --trace_cuda_only <bool> # gives CUDA only trace, default=False
    --power <int> # specifies s.t. input size = 2**power, default sizes are set to 2**16 to 2**24.
    --mx4_only <bool> # returns benchmark for mx4 only, default=False
    --fp8_only <bool> # returns benchmark for fp8_rowwise only, default=False

Other supported arguments:

--flush_gpu_cache_size_mb <int> # default=0
--iters <int> # default = 100
--group_size <int> # default = 32
--warmup_runs <int> # default = 10
--is_fwd <bool> # default = True

Differential Revision: D58172381

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D58172381

[netlify[bot]]

✅ Deploy Preview for pytorch-fbgemm-docs ready!

Name	Link
🔨 Latest commit	94b828c9256de1f41a0ff22c9cb2d919b12b9ea3
🔍 Latest deploy log	https://app.netlify.com/sites/pytorch-fbgemm-docs/deploys/6675d3e3b1a999000843f7d7
😎 Deploy Preview	https://deploy-preview-2759--pytorch-fbgemm-docs.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify site configuration.

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D58172381

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D58172381

[facebook-github-bot]

This pull request has been merged in pytorch/FBGEMM@918459ab1f0e92d35e8a12fe90498343c635e6d4.