[OpenMP] Use __OPENMP_NVPTX__ instead of _OPENMP in complex wrapper headers.

This is very similar to 7f1e6fc, just fixing a left-over.
With this, it should be possible to use both, -x cuda and -fopenmp in the same invocation,
enabling to use both OpenMP, targeting CPU, and CUDA, targeting the GPU.

Reviewed By: jdoerfert

Differential Revision: https://reviews.llvm.org/D90415

(cherry picked from commit eaee608)

Author: fodinabor

clang/lib/Headers/__clang_cuda_complex_builtins.h

@@ -16,7 +16,7 @@
 // to work with CUDA and OpenMP target offloading [in C and C++ mode].)
 
 #pragma push_macro("__DEVICE__")
-#ifdef _OPENMP
+#ifdef __OPENMP_NVPTX__
 #pragma omp declare target
 #define __DEVICE__ __attribute__((noinline, nothrow, cold, weak))
 #else
@@ -26,7 +26,7 @@
 // To make the algorithms available for C and C++ in CUDA and OpenMP we select
 // different but equivalent function versions. TODO: For OpenMP we currently
 // select the native builtins as the overload support for templates is lacking.
-#if !defined(_OPENMP)
+#if !defined(__OPENMP_NVPTX__)
 #define _ISNANd std::isnan
 #define _ISNANf std::isnan
 #define _ISINFd std::isinf
@@ -250,7 +250,7 @@ __DEVICE__ float _Complex __divsc3(float __a, float __b, float __c, float __d) {
 #undef _LOGBd
 #undef _LOGBf
 
-#ifdef _OPENMP
+#ifdef __OPENMP_NVPTX__
 #pragma omp end declare target
 #endif
 

clang/lib/Headers/openmp_wrappers/complex

@@ -18,7 +18,9 @@
 #include <cmath>
 
 #define __CUDA__
+#define __OPENMP_NVPTX__
 #include <__clang_cuda_complex_builtins.h>
+#undef __OPENMP_NVPTX__
 #endif
 
 // Grab the host header too.

clang/lib/Headers/openmp_wrappers/complex.h

@@ -18,7 +18,9 @@
 #include <math.h>
 
 #define __CUDA__
+#define __OPENMP_NVPTX__
 #include <__clang_cuda_complex_builtins.h>
+#undef __OPENMP_NVPTX__
 #endif
 
 // Grab the host header too.