Commit 9fcc4d8a authored by rusty1s's avatar rusty1s
Browse files

[skip ci]

parent 7134d457
...@@ -10,7 +10,8 @@ jobs: ...@@ -10,7 +10,8 @@ jobs:
strategy: strategy:
fail-fast: false fail-fast: false
matrix: matrix:
os: [ubuntu-16.04, macos-10.15, windows-latest] # os: [ubuntu-16.04, macos-10.15, windows-latest]
os: [windows-latest]
python-version: [3.6, 3.7, 3.8, 3.9] python-version: [3.6, 3.7, 3.8, 3.9]
torch-version: [1.8.0, 1.9.0] torch-version: [1.8.0, 1.9.0]
cuda-version: ['cpu', 'cu101', 'cu102', 'cu111'] cuda-version: ['cpu', 'cu101', 'cu102', 'cu111']
......
...@@ -6,11 +6,13 @@ ...@@ -6,11 +6,13 @@
AT_ASSERTM(x.device().is_cuda(), #x " must be CUDA tensor") AT_ASSERTM(x.device().is_cuda(), #x " must be CUDA tensor")
#define CHECK_INPUT(x) AT_ASSERTM(x, "Input mismatch") #define CHECK_INPUT(x) AT_ASSERTM(x, "Input mismatch")
#ifndef _WIN32
__device__ __inline__ at::Half __shfl_sync(const unsigned mask, __device__ __inline__ at::Half __shfl_sync(const unsigned mask,
const at::Half var, const at::Half var,
const unsigned int srcLane) { const unsigned int srcLane) {
return __shfl_sync(mask, (__half)var, srcLane); return __shfl_sync(mask, (__half)var, srcLane);
} }
#endif
__device__ __inline__ at::Half __shfl_down_sync(const unsigned mask, __device__ __inline__ at::Half __shfl_down_sync(const unsigned mask,
const at::Half var, const at::Half var,
......
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