Add the f16 sum kernel (fix).

This commit is contained in:
laurent
2023-06-26 22:02:22 +01:00
parent 93e24f29f4
commit 53fdbda683

View File

@ -44,7 +44,7 @@ extern "C" __global__ void FN_NAME( \
} \ } \
#if __CUDA_ARCH__ >= 530 #if __CUDA_ARCH__ >= 530
SUM_OP(float, sum_f32) SUM_OP(__half, sum_f16)
#endif #endif
SUM_OP(float, sum_f32) SUM_OP(float, sum_f32)