mirror of
https://github.com/ROCm/composable_kernel.git
synced 2026-05-01 20:21:23 +00:00
4208e28988
* This change introduces new pipelines with Intrawave scheduler and block gemm primitives that loads the scale tensor to registers to perform dequantization post MFMA on C tensor in registers. Scale tensor data, BQ is spliced across threads in registers and not stored in LDS. Current support is for the following combinations, but it should be fairly straightforward to extend support to more formats. fp8, fp8 -> f32 bf8, bf8 -> f32 fp8, i4 -> f32 bf8, i4 -> f32 Group size can go down to as low as K length of underlying WarpGemm primitive. * Solve merge conflict * [CK TILE] Update CHANGELOG.md --------- Co-authored-by: Vijay Krishnamoorthy <vjkrish@fb.com> Co-authored-by: ThomasNing <thomas.ning@amd.com> Co-authored-by: Cong Ma <congma13@amd.com>
1.4 KiB
1.4 KiB
GEMM Matrix Multiplication
This folder contains example for Block Scale GEMM using ck_tile tile-programming implementation.
build
# in the root of ck_tile
mkdir build && cd build
# you can replace <arch> with the appropriate architecture (for example gfx90a or gfx942) or leave it blank
../script/cmake-ck-dev.sh ../ <arch>
# The aquant pipeline method on the gemm calculation
make tile_example_gemm_aquant_basic -j
make tile_example_gemm_bquant_basic -j
This will result in an executable build/bin/tile_example_gemm_aquant_basic
example
args:
-b batch size (default:1)
-m m dimension (default:1024)
-n n dimension (default:2048)
-k k dimension (default:64)
-a_layout Tensor A data layout (default: R)
-b_layout Tensor B data layout (default: R)
-c_layout Tensor C data layout (default: R)
-stride_a Tensor A stride (default:0)
-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:2)
-e Absolute error tolerance (default:1e-5)
-prec data type. fp16/bf16/fp8/bf8/int8 (default:fp16)
-warmup number of iterations before benchmark the kernel (default:10)
-repeat number of iterations to benchmark the kernel (default:100)
-timer gpu:gpu timer, cpu:cpu timer (default:gpu)