From b9d3a22d31f8ecccc2abe26789f9e6a3490f08aa Mon Sep 17 00:00:00 2001 From: carlushuang Date: Thu, 31 Dec 2020 14:37:33 +0800 Subject: [PATCH] update readme (#73) --- README.md | 12 ++++++++++-- 1 file changed, 10 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index 391d52cc..7d811a55 100644 --- a/README.md +++ b/README.md @@ -7,9 +7,17 @@ Code generator for implicit gemm algorithm (generic tensor contraction) # Generate kernel since f-string is utilized in python, require python >= 3.6 to run. ``` -# generate code based on tunable configuration -python3 igemm_codegen.py config/igemm_bwd_gtc.config +# generate code based on tunable configuration, use one of following command to generate each direction +python3 igemm_codegen.py config/igemm_fwd_gtc_gfx908.config +python3 igemm_codegen.py config/igemm_bwd_gtc_gfx908.config +python3 igemm_codegen.py config/igemm_wrw_gtc_gfx908.config + +# or auto generate code for all possible combinations, use one of following command to generate each direction +python3 igemm_codegen.py config/igemm_fwd_gtc_gfx908_seq.config +python3 igemm_codegen.py config/igemm_bwd_gtc_gfx908_seq.config +python3 igemm_codegen.py config/igemm_wrw_gtc_gfx908_seq.config ``` + The output file will result in `out` directory. result in a assembly file `*.s` and several `*.inc` for different tile size, a codeobject `*.hsaco` and a host driver executable `conv_driver.exe`. This executable accept same cmdline argument as [MIOpenDriver](https://rocmsoftwareplatform.github.io/MIOpen/doc/html/driver.html). e.g. ``` ./conv_driver.exe conv -n 128 -c 1024 -H 17 -W 17 -k 1024 -y 1 -x 1 -p 0 -q 0 -u 1 -v 1 -l 1 -j 1 -F 2 -V 1