* Update splitk_gemm.cu

* Update gemm_bias_relu.cu

* Update mma_sm75.h
This commit is contained in:
hwu36
2020-07-13 14:25:52 -07:00
committed by GitHub
parent fd7e058d0c
commit 4dac7490e6
3 changed files with 3 additions and 4 deletions
+1 -1
View File
@@ -205,7 +205,7 @@ int run() {
cutlass::HostTensor<ElementInputA, LayoutInputA> tensor_a(
problem_size.mk()); // <- Create matrix A with dimensions M x K
cutlass::HostTensor<ElementInputB, LayoutInputB> tensor_b(
problem_size.nk()); // <- Create matrix B with dimensions N x K
problem_size.kn()); // <- Create matrix B with dimensions K x N
cutlass::HostTensor<ElementOutput, LayoutOutput> tensor_c(
problem_size.mn()); // <- Create matrix C with dimensions M x N
cutlass::HostTensor<ElementOutput, LayoutOutput> tensor_d(
+1 -2
View File
@@ -132,7 +132,7 @@ int run() {
cutlass::HostTensor<ElementInputA, LayoutInputA> tensor_a(
problem_size.mk()); // <- Create matrix A with dimensions M x K
cutlass::HostTensor<ElementInputB, LayoutInputB> tensor_b(
problem_size.nk()); // <- Create matrix B with dimensions N x K
problem_size.kn()); // <- Create matrix B with dimensions K x N
cutlass::HostTensor<ElementOutput, LayoutOutput> tensor_c_bias(
{problem_size.m(), 1}); // <- Create matrix C with dimensions M x 1
@@ -234,7 +234,6 @@ int run() {
tensor_a.device_ref(),
tensor_b.device_ref(),
0,
tensor_c_bias.device_ref(),
tensor_ref_d.device_ref());
// Wait for kernels to finish