[llvm] [AMDGPU] Add flag to enable expensive trip counts for runtime unroll. (PR #171735)

[via llvm-commits]

================
@@ -0,0 +1,77 @@
+; RUN: opt -mtriple=amdgcn-amd-amdhsa -passes=loop-unroll -S %s -o - | FileCheck -check-prefixes=CHECK-NOUNROLL %s
+; RUN: opt -amdgpu-unroll-runtime-expensive-trip-count -mtriple=amdgcn-amd-amdhsa -passes=loop-unroll -S %s -o - | FileCheck -check-prefixes=CHECK-UNROLL %s
----------------
carlobertolli wrote:

I usually prefer minimal checks if the output is simple enough. There's no chance we are going to get more store instructions on the output given there is a single one in the input, unless some form of unrolling happens.

https://github.com/llvm/llvm-project/pull/171735