ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-11 17:00:18 +00:00
Code Issues Packages Projects Releases Wiki Activity

retune implicit gemm v4r1

Browse Source
This commit is contained in:
Chao Liu
2019-07-30 12:10:28 -05:00
parent cd8de11218
commit c5e5a9307b
2 changed files with 39 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
driver/src/driver.cpp
View File

@@ -72,11 +72,11 @@ int main(int argc, char* argv[])
using namespace ck;
#if 0
constexpr index_t N = 2;
constexpr index_t C = 16;
constexpr index_t N = 256;
constexpr index_t C = 1536;
constexpr index_t HI = 8;
constexpr index_t WI = 8;
constexpr index_t K = 128;
constexpr index_t K = 512;
constexpr index_t Y = 1;
constexpr index_t X = 1;
@@ -532,7 +532,7 @@ int main(int argc, char* argv[])
#elif 0
device_convolution_implicit_gemm_v3_nchw_cyxk_nkhw(
(in_nchw_desc, in_nchw, wei_kcyx_desc, wei_kcyx, out_nkhw_desc, out_nkhw_device, nrepeat);
#elif 0
#elif 1
device_convolution_implicit_gemm_v4r1_nchw_kcyx_nkhw(in_nchw_desc,
in_nchw,
wei_kcyx_desc,