[llvm] [NewGVN] Check intrinsic ID before accessing operands (PR #141571)

[Mariusz Sikora via llvm-commits]

================
@@ -0,0 +1,17 @@
+; NOTE: Assertions have been autogenerated by utils/update_test_checks.py UTC_ARGS: --version 5
+; RUN: opt -S -passes=newgvn < %s | FileCheck %s
----------------
mariusz-sikora-at-amd wrote:

OK, I will move this test and I will add `REQUIRES: amdgpu-registered-target` since I'm using here `llvm.amdgcn.s.barrier`

https://github.com/llvm/llvm-project/pull/141571