mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-21 05:19:20 +00:00
[CK_TILE] Add depthwise conv2d forward kernel (FP16/FP32) (#6838) ## Motivation CK currently has no kernel optimized for depthwise convolution (G=C_in=C_out, C=K=1 per group) and existing generic paths perform poorly for this workload. This PR adds a dedicated depthwise conv forward kernel in CK Tile. ## Technical Details Adds a dedicated depthwise conv2d forward op to CK Tile that performs direct convolution rather than falling back to the generic GEMM path. The kernel is templatized by filter size, stride, and data type, and compiled into ~60 instances covering common configurations (kernel 3/5/7/9, stride 1/2, FP16/FP32). Supports both CDNA (gfx942/gfx950) and RDNA (gfx1100/gfx1200) architectures. ## Test Plan - [x] Correctness and performance validated on gfx942, gfx950, and gfx1100, with ckProfiler `grouped_conv_fwd` as baseline. - [ ] MI300A (gfx942) and gfx1200 validation. ## Submission Checklist - [x ] Look over the contributing guidelines at https://github.com/ROCm/ROCm/blob/develop/CONTRIBUTING.md#pull-requests. AICK-1137
common
this folder is designed not to be included directly by use, e.g. if use include ck_tile/ops/fmha.hpp, then everything under common should also be included.
to achieve this we will duplicate the header include path under common to other module under ops/* inside remod.py. for internal developer, you can also include ck_tile/ops/common.hpp for convenience. (and so does external users...)