bug #1526 - CUDA compilation fails on CUDA 9.x SDK when arch is set to compute_60 and/or above

This commit is contained in:
Jiandong Ruan 2018-09-08 12:05:33 -07:00
parent edfb7962fd
commit 6dcd2642aa

View File

@ -272,7 +272,11 @@ namespace half_impl {
// conversion steps back and forth.
EIGEN_STRONG_INLINE __device__ half operator + (const half& a, const half& b) {
#if defined(EIGEN_CUDACC_VER) && EIGEN_CUDACC_VER >= 90000
return __hadd(::__half(a), ::__half(b));
#else
return __hadd(a, b);
#endif
}
EIGEN_STRONG_INLINE __device__ half operator * (const half& a, const half& b) {
return __hmul(a, b);
@ -281,9 +285,13 @@ EIGEN_STRONG_INLINE __device__ half operator - (const half& a, const half& b) {
return __hsub(a, b);
}
EIGEN_STRONG_INLINE __device__ half operator / (const half& a, const half& b) {
#if defined(EIGEN_CUDACC_VER) && EIGEN_CUDACC_VER >= 90000
return __hdiv(a, b);
#else
float num = __half2float(a);
float denom = __half2float(b);
return __float2half(num / denom);
#endif
}
EIGEN_STRONG_INLINE __device__ half operator - (const half& a) {
return __hneg(a);