ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-14 02:02:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Restructure gridwise and blockwise GEMM, add tensor contraction and FWD-v4r5 (#36)

Browse Source 
* experimenting magic number division

* overhauling fwd-v4r4 to clearly reflect transformation graph

* added fwd-v4r5

* bug fix for make_dynamic_naive_tensor_descriptor_aligned_v2

* bug fix and added sanity-check in transform_dynamic_tensor_descriptor

* added conv_driver_v2

[ROCm/composable_kernel commit: 30072aec37]
This commit is contained in:
Chao Liu
2021-06-09 23:53:08 -05:00
committed by GitHub
parent 040023fdcd
commit f4acec502e
38 changed files with 4791 additions and 2050 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
script/run.sh Executable file
View File

@@ -0,0 +1,17 @@
#!/bin/bash
#make -j conv_driver
make -j conv_driver_v2
LAYOUT=$1
ALGO=$2
VERIFY=$3
INIT=$4
LOG=$5
REPEAT=$6
###################### layout algo verify init log repeat N__ K__ C__ Y X Hi_ Wi__ Strides Dilations LeftPads RightPads
driver/conv_driver_v2 $LAYOUT $ALGO $VERIFY $INIT $LOG $REPEAT 128 128 192 3 3 71 71 2 2 1 1 1 1 1 1
#driver/conv_driver_v2 $LAYOUT $ALGO $VERIFY $INIT $LOG $REPEAT 128 384 192 3 3 35 35 2 2 1 1 0 0 0 0
#driver/conv_driver_v2 $LAYOUT $ALGO $VERIFY $INIT $LOG $REPEAT 128 128 128 1 7 17 17 1 1 1 1 0 3 0 3
#driver/conv_driver_v2 $LAYOUT $ALGO $VERIFY $INIT $LOG $REPEAT 128 256 256 3 3 14 14 1 1 1 1 1 1 1 1