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

tidy up

Browse Source
This commit is contained in:
Chao Liu
2019-04-08 10:48:29 -05:00
parent c9fa46af0b
commit 268d1c717c
14 changed files with 120 additions and 102 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
driver/driver.hip.cpp
View File

@@ -353,7 +353,7 @@ void host_winograd_3x3_convolution(const Tensor<TIn>& in_nchw,
std::size_t ho = HoPerTile * htile + j;
for(int i = 0; i < WoPerTile; ++i)
{
std::size_t wo = WoPerTile * wtile + i;
std::size_t wo = WoPerTile * wtile + i;
out_nkhw(n, k, ho, wo) = out_hold(n, k, htile, wtile, j, i);
}
}
@@ -409,7 +409,7 @@ int main(int argc, char* argv[])
constexpr index_t HPad = 0;
constexpr index_t WPad = 0;
#elif 1
#elif 0
// 3x3, 34x34
constexpr index_t N = 64;
constexpr index_t C = 256;
@@ -580,7 +580,7 @@ int main(int argc, char* argv[])
constexpr index_t HPad = 0;
constexpr index_t WPad = 0;
#elif 0
#elif 1
// 1x1 filter, 14x14 image, C = 2048
constexpr index_t N = 128;
constexpr index_t C = 2048;
@@ -661,7 +661,7 @@ int main(int argc, char* argv[])
device_direct_convolution_2_nchw_kcyx_nkhw
#elif 0
device_direct_convolution_2_vectorized_nchw_kcyx_nkhw
#elif 1
#elif 0
device_implicit_gemm_convolution_1_chwn_cyxk_khwn
#elif 1
device_implicit_gemm_convolution_2_chwn_cyxk_khwn