[clang] [llvm] [NVPTX] Add support for PTX 8.6 and CUDA 12.6 (12.8) (PR #123398)

[Sergey Kozub via llvm-commits]

================
@@ -682,6 +688,9 @@ void NVPTX::getNVPTXTargetFeatures(const Driver &D, const llvm::Triple &Triple,
   case CudaVersion::CUDA_##CUDA_VER:                                           \
     PtxFeature = "+ptx" #PTX_VER;                                              \
     break;
+    CASE_CUDA_VERSION(129, 86);
+    CASE_CUDA_VERSION(128, 86);
+    CASE_CUDA_VERSION(127, 85);
----------------
sergey-kozub wrote:

Updated, thank you.
Also, changed line 235 in "clang/lib/Basic/Cuda.cpp", as PTX8.6 supports sm100a.

https://github.com/llvm/llvm-project/pull/123398