ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-01 20:21:23 +00:00
Code Issues Packages Projects Releases Wiki Activity

Bf16*fp4 gemm (#2801)

Browse Source 
* support bf16*mxfp4 gemm

* rebase bf16*fp4 example to develop branch

* Clean up commented debug code in GEMM kernel

* rename example folder

* support bf16*mxfp4 gemm

* rebase bf16*fp4 example to develop branch

* Clean up commented debug code in GEMM kernel

* rename example folder

* rebase to new develop

* fix clang format

* update code according to reviewer's comment

* Update README.md

* update code according to reviewer's comment

* update code according to reviewer's comment

* Update CMakeLists.txt

* Update README.md

* Update CMakeLists.txt

* Delete files

* Delete files

* Add unit tests

* Update test_gemm_quant_base.hpp

* merge bf16*fp4 example to develop branch

* fix clang format

* fix clang format

* Update CMakeLists.txt

* fix ci test

* fix clang format

* resolve conflicts

---------

Co-authored-by: eliotwang <charyang@smci355-ccs-aus-m10-29.cs-aus.dcgpu>
Co-authored-by: ShaoChunLee <Shao-Chun.Lee@amd.com>
Co-authored-by: Illia Silin <98187287+illsilin@users.noreply.github.com>
Co-authored-by: illsilin_amdeng <Illia.Silin@amd.com>
Co-authored-by: Thomas Ning <Thomas.Ning@amd.com>
This commit is contained in:
eliotwang
2025-12-11 23:20:29 +08:00
committed by GitHub
parent ce99cab605
commit 715671e419
23 changed files with 1260 additions and 137 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
example/ck_tile/38_block_scale_gemm/README.md
View File

@@ -23,7 +23,7 @@ This folder contains examples of quant GEMMs using the ck_tile tile-programming
- **Preshuffled GEMM**: Shuffle the GEMM of B (weight) matrix in the warp layout and bypass the shared memory to do the GEMM calculation. Best performance solution for GEMM.
- **TransposeC**: Transpose the C Matrix Output layout to have the best coalesced scale reading
- **Preshuffled Quant**: Preshuffle the input matrix to load multiple Quant warp blocks along the selected dimension.
- **Precision**: Supports fp16, bf16, fp8, bf8, int4 (for B Matrix).
- **Precision**: Supports fp16, bf16, fp8, bf8, int4 (for B Matrix), uint8 (split into two fp4 in the pipeline (for B Matrix)).
- **Validation**: CPU/GPU validation and error tolerance options.
## build
@@ -53,7 +53,7 @@ args:
-stride_b Tensor B stride (default:0)
-stride_c Tensor C stride (default:0)
-v 0: No validation, 1: Validation on CPU, 2: Validation on GPU (default:1)
-prec Data type. For AQuant: fp8, bf8, i4fp8, or i4bf8; for Bquant: fp8, bf8, fp8i4, or bf8i4 (default for both AQuant and Bquant: fp8)
-prec Data type. For AQuant: fp8, bf8, i4fp8, or i4bf8; for Bquant: fp8, bf8, fp8i4, bf8i4, or bf16fp4 (default for both AQuant and Bquant: fp8)
-warmup Number of iterations before benchmarking the kernel (default:50)
-repeat Number of iterations to benchmark the kernel (default:1000)
-timer gpu:gpu timer, cpu:cpu timer (default:gpu)