[Openmp-commits] [openmp] [OpenMP][AMDGPU] Adapt dynamic callstack sizes to HIP behavior (PR #74080)

[Johannes Doerfert via Openmp-commits]

================
@@ -3198,7 +3233,8 @@ Error AMDGPUKernelTy::launchImpl(GenericDeviceTy &GenericDevice,
 
   // Push the kernel launch into the stream.
   return Stream->pushKernelLaunch(*this, AllArgs, NumThreads, NumBlocks,
-                                  GroupSize, StackSize, ArgsMemoryManager);
+                                  GroupSize, static_cast<uint32_t>(StackSize),
----------------
jdoerfert wrote:

Change it to uint32_t everywhere instead. Let's assume we stop at 4GB stack size.

https://github.com/llvm/llvm-project/pull/74080