[llvm] [NVPTX] Add asynchronous store intrinsics (PR #200768)

[Durgadoss R via llvm-commits]

================
@@ -1116,10 +1116,13 @@ NVPTXTargetLowering::NVPTXTargetLowering(const NVPTXTargetMachine &TM,
                       MVT::v64f32, MVT::v128f32},
                      Custom);
 
-  // Custom lowering for tcgen05.st vector operands
+  // Custom lowering for tcgen05.st vector operands and the st.async mbarrier
----------------
durga4github wrote:

remove the `mbarrier` from here

https://github.com/llvm/llvm-project/pull/200768