[flang-commits] [flang] [flang][cuda] Introduce stream cast op (PR #136050)

Zhen Wang via flang-commits flang-commits at lists.llvm.org
Wed Apr 16 20:57:31 PDT 2025


================
@@ -370,4 +371,25 @@ def cuf_SharedMemoryOp
       CArg<"llvm::ArrayRef<mlir::NamedAttribute>", "{}">:$attributes)>];
 }
 
+def cuf_StreamCastOp : cuf_Op<"stream_cast", [NoMemoryEffect]> {
+  let summary = "Adapt a stream value to a GPU async token";
+
+  let description = [{
+    Cast a stream object reference as a GPU async token. This is useful to be
+    able to connect the stream representation of CUDA Fortran and the async
+    mechanism of the GPU dialect.
+    Later in the lowering this will becoming a no op.
----------------
wangzpgi wrote:

becoming -> become

https://github.com/llvm/llvm-project/pull/136050


More information about the flang-commits mailing list