[clang] [llvm] [AArch64][SME] Split FP8 FTMOPA intrinsics (PR #203310)

[via cfe-commits]

================
@@ -0,0 +1,42 @@
+; RUN: opt -S < %s | FileCheck %s
+; RUN: llvm-as %s -o - | llvm-dis | FileCheck %s
----------------
CarolineConcatto wrote:

 The  llvm-as performs the upgrade when it reads the old textual IR and writes bitcode. That was intentional for this test: the input uses the old intrinsic spelling, and the llvm-as | llvm-dis pipeline verifies that AutoUpgrade accepts it and  serializes the upgraded intrinsic name. The llvm-dis step is only there to make the upgraded bitcode visible to FileCheck, so we can check that the intrinsic was renamed correctly.
But I think you  want a .bc file using the old intrinsic like I see for the experimental.vector.insert intrinsics. I will replace for this type of test.

https://github.com/llvm/llvm-project/pull/203310