提交 f2b84d54 编写于 作者: L liaogang 提交者: GitHub

Merge pull request #15 from stoneyang/hl_device_func

fixed build issue of double definition of atomicAdd on modern GPUs
......@@ -16,6 +16,8 @@ limitations under the License. */
#ifndef HL_DEVICE_FUNCTIONS_CUH_
#define HL_DEVICE_FUNCTIONS_CUH_
#if defined(__CUDA_ARCH__) && __CUDA_ARCH__ < 600
namespace hppl {
static __inline__ __device__ double atomicAdd(double* address, double val) {
......@@ -38,4 +40,6 @@ static __inline__ __device__ double atomicAdd(double* address, double val) {
using hppl::atomicAdd;
#endif
#endif /* HL_DEVICE_FUNCTIONS_CUH_ */
Markdown is supported
0% .
You are about to add 0 people to the discussion. Proceed with caution.
先完成此消息的编辑!
想要评论请 注册