Skip to content

GitLab

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