ROCm/composable_kernel
1
0
Fork 0
mirror of https://github.com/ROCm/composable_kernel.git synced 2026-05-12 01:10:17 +00:00
Code Issues Packages Projects Releases Wiki Activity

adding implicit gemm v3

Browse Source
This commit is contained in:
Chao Liu
2019-05-23 22:10:40 -05:00
parent 8a4b59785b
commit 1cc683a3a3
16 changed files with 347 additions and 95 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
driver/driver.hip.cpp
View File

@@ -411,7 +411,18 @@ void check_error(const Tensor<T>& ref, const Tensor<T>& result)
int main(int argc, char* argv[])
{
#if 1
#if 0
constexpr index_t N = 8;
constexpr index_t C = 8;
constexpr index_t HI = 3;
constexpr index_t WI = 18;
constexpr index_t K = 128;
constexpr index_t Y = 3;
constexpr index_t X = 3;
constexpr index_t HPad = 0;
constexpr index_t WPad = 0;
#elif 1
// 3x3, 34x34
constexpr index_t N = 64;
constexpr index_t C = 256;
@@ -635,11 +646,13 @@ int main(int argc, char* argv[])
if(do_verification)
{
#if 1
if(Y == 3 && X == 3)
{
host_winograd_3x3_convolution(in_nchw, wei_kcyx, out_nkhw_host, lower_pads, upper_pads);
}
else
#endif
{
host_direct_convolution(in_nchw, wei_kcyx, out_nkhw_host, lower_pads, upper_pads);
}