[flang] [llvm] [Flang][OpenMP][Runtime] Minor Flang runtime for OpenMP AMDGPU (PR #152631)

[Peter Klausler via llvm-commits]

================
@@ -23,6 +23,16 @@
 #define STD_FILL_N_UNSUPPORTED 1
 #endif
 
+#if !defined(STD_MEMSET_UNSUPPORTED) && \
+    (defined(__CUDACC__) || defined(__CUDA__)) && defined(__CUDA_ARCH__)
+#define STD_MEMSET_UNSUPPORTED 1
----------------
klausler wrote:

Won't this change existing compilations that work with CUDA?

https://github.com/llvm/llvm-project/pull/152631