[llvm] [AArch64] Parameterize repeated FP divisor combine threshold by subtarget (PR #216930)

[Sander de Smalen via llvm-commits]

================
@@ -1,6 +1,8 @@
 ; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
-; RUN: llc -mtriple=aarch64-unknown-unknown < %s | FileCheck %s --check-prefixes=CHECK,CHECK-SD
-; RUN: llc -mtriple=aarch64-unknown-unknown < %s -global-isel -global-isel-abort=2 2>&1 | FileCheck %s --check-prefixes=CHECK,CHECK-GI
+; RUN: llc -mtriple=aarch64-unknown-unknown < %s | FileCheck %s --check-prefixes=CHECK,CHECK-THREE,CHECK-SDGI,CHECK-SD
+; RUN: llc -mtriple=aarch64-unknown-unknown < %s -global-isel -global-isel-abort=2 2>&1 | FileCheck %s --check-prefixes=CHECK,CHECK-THREE,CHECK-SDGI,CHECK-GI
+; RUN: llc -mtriple=aarch64-unknown-unknown -mcpu=neoverse-v2 < %s | FileCheck %s --check-prefixes=CHECK,CHECK-THREE,CHECK-MCPU
+; RUN: llc -mtriple=aarch64-unknown-unknown -mcpu=a64fx < %s | FileCheck %s --check-prefixes=CHECK,CHECK-MCPU,CHECK-TWO
----------------
sdesmalen-arm wrote:

Rather than adding a new RUN line  (which makes things a bit complicated with all the CHECk prefixes), I would suggest duplicating one of the tests that has 2 fdivs and adding the `"target-cpu"="a64fx"` attribute to it.

https://github.com/llvm/llvm-project/pull/216930