tra added inline comments.
================
Comment at: clang/lib/Headers/__clang_cuda_cmath.h:81-90
+__DEVICE__ bool isinf(long double __x) { return ::__isinfl(__x); }
__DEVICE__ bool isfinite(float __x) { return ::__finitef(__x); }
// For inscrutable reasons, __finite(), the double-precision version of
// __finitef, does not exist when compiling for MacOS. __isfinited is
available
// everywhere and is just as good.
__DEVICE__ bool isfinite(double __x) { return ::__isfinited(__x); }
+__DEVICE__ bool isfinite(long double __x) { return ::__finitel(__x); }
----------------
CUDA [[
https://docs.nvidia.com/cuda/cuda-c-programming-guide/index.html#long-double |
does not support]] long double on device side, so I would rather see a
compilation error when someone attempts to use a long double, as opposed to
providing inconsistent support for it.
I wonder if we could just provide declarations for these functions. This should
allow math_functions.hpp to compile, but would still prevent any use of `long
double` on device side.
Repository:
rC Clang
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D60220/new/
https://reviews.llvm.org/D60220
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
