[llvm] [AMDGPU] Precommit Tests for 'Fold redundant inf/nan checks into frexp instructions' (PR #214964)

[Krzysztof Drewniak via llvm-commits]

================
@@ -0,0 +1,1130 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -mtriple=amdgpu9.00--amdpal < %s | FileCheck -check-prefix=GFX9 %s
+; RUN: llc -mtriple=amdgpu10.30--amdpal < %s | FileCheck -check-prefix=GFX10 %s
+; RUN: llc -mtriple=amdgpu11.00--amdpal < %s | FileCheck -check-prefix=GFX11 %s
+; RUN: llc -mtriple=amdgpu9.50--amdpal < %s | FileCheck -check-prefix=GFX950 %s
+; RUN: llc -mtriple=amdgpu6.00--amdpal < %s | FileCheck -check-prefix=GFX6 %s
----------------
krzysz00 wrote:

Why do we have `amdpal` here? Also, these are in a slightly weird order?

But that's nitpicking and isn't important to getting this landed

https://github.com/llvm/llvm-project/pull/214964