Commit c39c573e authored by Chao Liu's avatar Chao Liu
Browse files

refactor

parent c9af4dec
......@@ -102,8 +102,8 @@ void device_implicit_gemm_convolution_2_cnhw_srck_knhw(InDesc,
constexpr unsigned KPerThread = 8;
constexpr unsigned CPerThread = 1;
constexpr unsigned GemmThreadPerClusterRow = 4;
constexpr unsigned GemmThreadPerClusterColumn = 4;
constexpr unsigned GemmRowThreadPerCluster = 4;
constexpr unsigned GemmColumnThreadPerCluster = 4;
constexpr unsigned InBlockCopyThreadPerDim0 = 2;
constexpr unsigned InBlockCopyThreadPerDim1 = 64;
......@@ -149,8 +149,8 @@ void device_implicit_gemm_convolution_2_cnhw_srck_knhw(InDesc,
BPerThread,
KPerThread,
CPerThread,
GemmThreadPerClusterRow,
GemmThreadPerClusterColumn,
GemmRowThreadPerCluster,
GemmColumnThreadPerCluster,
InBlockCopyThreadPerDim0,
InBlockCopyThreadPerDim1>
<<<grid_dim, block_dim>>>(in_cnhw_desc,
......
......@@ -20,8 +20,8 @@ template <unsigned GridSize,
unsigned BPerThread,
unsigned KPerThread,
unsigned CPerThread,
unsigned GemmThreadPerClusterRow,
unsigned GemmThreadPerClusterColumn,
unsigned GemmRowThreadPerCluster,
unsigned GemmColumnThreadPerCluster,
unsigned InBlockCopyThreadPerDim0,
unsigned InBlockCopyThreadPerDim1>
__global__ void
......@@ -177,8 +177,8 @@ gridwise_implicit_gemm_convolution_3_cnhw_srck_knhw(InGlobalDesc,
false,
false,
CPerThread,
GemmThreadPerClusterRow,
GemmThreadPerClusterColumn,
GemmRowThreadPerCluster,
GemmColumnThreadPerCluster,
true>{};
// LDS
......
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