Skip to content

[Feature]: Intra-kernel profiling #831

Description

@benenzhu

Suggestion Description

Intra-kenrel profiling is really useful for communications kernels.
Also it is useful for compute-bound kernels too, to identify the XCD-imbanlance, and tail effect of some grids.

Some reference for other dsl.

  1. cutedsl: https://docs.nvidia.com/cutlass/latest/media/docs/pythonDSL/cute_dsl_general/iket_profiling.html
  2. tilelang: [Feature] Support iket profiler for CUDA backend tile-ai/tilelang#2515
  3. triton: https://github.com/triton-lang/triton/blob/main/third_party/proton/tutorials/intra_kernel/example_dsl.py
  4. flashinfer: feat: flashinfer intra-kernel profiler flashinfer-ai/flashinfer#913

Also here is a blog for how to use it to identify uneven work distribution in MI300x intra-node all2all implementations:
https://gau-nernst.github.io/amd-a2a/#intra-kernel-profiling

Image

Operating System

No response

GPU

No response

ROCm Component

No response

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Fields

    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions