Skip to content

Conversation

@bartekxk
Copy link
Contributor

@bartekxk bartekxk commented Nov 3, 2025

Proposed changes

Refactor Conv configs and Conv Elementwise. Change GemmConfigs to ConvConfigs and remove Elementwise from Conv Traits. Remove not needed members

Checklist

Please put an x into the boxes that apply. You can also fill these out after creating the PR. If you're not sure, please don't hesitate to ask.

  • I have added tests relevant to the introduced functionality, and the unit tests are passing locally
  • I have added the test to REGRESSION_TESTS list defined at the top of CMakeLists.txt in tests/CMakeLists.txt, IF the test takes more than 30 seconds to run.
  • I have added inline documentation which enables the maintainers with understanding the motivation
  • I have removed the stale documentation which is no longer relevant after this pull request
  • (If this change is user-facing) I have added release notes which provide the end users with a brief summary of the improvement from this pull request
  • I have run clang-format on all changed files
  • Any dependent changes have been merged

Discussion

If this is a relatively large or complex change, feel free to start a discussion by explaining why you chose the solution you did and what alternatives you considered

Copy link
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR refactors the grouped convolution code to improve API design by removing the CDElementwise template parameter from GroupedConvTraits and moving it to the kernel arguments level. Additionally, it renames GemmConfig to ConvConfig throughout the codebase for better clarity, and removes unused configuration fields (PermuteA, PermuteB, UseStructuredSparsity, Preshuffle, TiledMMAPermuteN).

  • Moved CDElementwise from GroupedConvTraits to GroupedConvFwdKernelArgs template parameters
  • Renamed GemmConfig to ConvConfig across all example files
  • Removed unused configuration parameters and simplified TileGemmShape and TileGemmUniversalTraits instantiations

Reviewed Changes

Copilot reviewed 14 out of 14 changed files in this pull request and generated 1 comment.

Show a summary per file
File Description
include/ck_tile/ops/grouped_convolution/utils/grouped_convolution_utils.hpp Removed CDElementwise_ template parameter from GroupedConvTraits
include/ck_tile/ops/grouped_convolution/kernel/grouped_convolution_forward_kernel.hpp Added CDElementwise_ to GroupedConvFwdKernelArgs template parameters
example/ck_tile/20_grouped_convolution/run_grouped_convolution_fwd_example.inc Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/run_grouped_convolution_bwd_data_example.inc Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/grouped_convolution_forward_large_tensor_invoker.hpp Renamed GemmConfig to ConvConfig, removed unused parameters, added duplicate comment
example/ck_tile/20_grouped_convolution/grouped_convolution_forward_large_tensor.cpp Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/grouped_convolution_forward_invoker.hpp Renamed GemmConfig to ConvConfig, removed unused parameters, simplified trait instantiations
example/ck_tile/20_grouped_convolution/grouped_convolution_forward_bias_clamp.cpp Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/grouped_convolution_forward.cpp Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/grouped_convolution_backward_weight_two_stage_invoker.hpp Removed unused parameters from TileGemmShape instantiation
example/ck_tile/20_grouped_convolution/grouped_convolution_backward_weight_invoker.hpp Removed unused parameters from TileGemmShape instantiation
example/ck_tile/20_grouped_convolution/grouped_convolution_backward_data_invoker.hpp Renamed GemmConfig to ConvConfig, removed unused parameters
example/ck_tile/20_grouped_convolution/grouped_convolution_backward_data.cpp Renamed GemmConfig to ConvConfig
example/ck_tile/20_grouped_convolution/conv_configs.hpp Removed unused configuration fields

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

auto kargs = Kernel::MakeKernelArgs(args);

// Populate split-image metadata ONLY if using split-image kernel
// Populate split-image metadata ONLY if using split-image kernel
Copy link

Copilot AI Nov 3, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Duplicate comment on consecutive lines. Remove one of the duplicate comments.

Suggested change
// Populate split-image metadata ONLY if using split-image kernel

Copilot uses AI. Check for mistakes.
@bartekxk bartekxk merged commit 8681ced into develop Nov 4, 2025
49 checks passed
@bartekxk bartekxk deleted the barkocot/refactor-conv-tile branch November 4, 2025 14:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants