Skip to content

GitLab

You need to sign in or sign up before continuing.
Commit 85c1ff1c authored by Chao Liu's avatar Chao Liu
Browse files

change perf config for debuggging

parent 23c626a9
Hide whitespace changes
Inline Side-by-side
Showing with 2 additions and 2 deletions
driver/device_implicit_gemm_convolution_2_chwn_cyxk_khwn.hpp
View file @ 85c1ff1c
...@@ -190,7 +190,7 @@ void device_implicit_gemm_convolution_2_chwn_cyxk_khwn(InDesc, ...@@ -190,7 +190,7 @@ void device_implicit_gemm_convolution_2_chwn_cyxk_khwn(InDesc,
constexpr index_t WeiBlockCopyDataPerRead = 4; constexpr index_t WeiBlockCopyDataPerRead = 4;
constexpr index_t BlockSize = 256; constexpr index_t BlockSize = 256;
#elif 1 #elif 0
// 1x1, 14x14, Vega 20, disable lds_double_buffer, enable register double buffer // 1x1, 14x14, Vega 20, disable lds_double_buffer, enable register double buffer
constexpr index_t BPerBlock = 64; constexpr index_t BPerBlock = 64;
constexpr index_t KPerBlock = 128; constexpr index_t KPerBlock = 128;
......
src/include/gridwise_implicit_gemm_convolution_2_chwn_cyxk_khwn.hip.hpp
View file @ 85c1ff1c
...@@ -315,7 +315,7 @@ class gridwise_implicit_gemm_convolution_2_chwn_cyxk_khwn ...@@ -315,7 +315,7 @@ class gridwise_implicit_gemm_convolution_2_chwn_cyxk_khwn
auto f_accum = [](auto& acc, const auto&& v) { acc += v; }; auto f_accum = [](auto& acc, const auto&& v) { acc += v; };
#if 0 #if 0
blockwise_gemm.Run blockwise_gemm.Run
#elif 1 #elif 0
blockwise_gemm.Run_RegisterDoubleBuffer blockwise_gemm.Run_RegisterDoubleBuffer
#elif 1 #elif 1
blockwise_gemm.Run_asm blockwise_gemm.Run_asm
......
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment