To make CUDA_LAUNCH_KERNEL_HELPER support large size.

test=develop
revert-15296-async_double_buffered_py_reader
guoshengCS 6 years ago
parent ba02ac4692
commit 5dfce93101

@ -54,6 +54,8 @@ inline static int RoundToPowerOfTwo(int dim) {
} break
#define CUDA_LAUNCH_KERNEL_HELPER(...) \
CUDA_LAUNCH_KERNEL_BASE(1024, ##__VA_ARGS__); \
CUDA_LAUNCH_KERNEL_BASE(512, ##__VA_ARGS__); \
CUDA_LAUNCH_KERNEL_BASE(256, ##__VA_ARGS__); \
CUDA_LAUNCH_KERNEL_BASE(128, ##__VA_ARGS__); \
CUDA_LAUNCH_KERNEL_BASE(64, ##__VA_ARGS__); \

Loading…
Cancel
Save