Compiler projects using llvm
# REQUIRES: amdgpu-registered-target
## Check that functions with different IR comments in the output are not merged

# Replace llc with cat, so we can simulate hypothetical output without actually running llc.
# Copy the simulated output to the temporary directory.
# RUN: cp -f %S/Inputs/amdgpu_no_merge_comments-O0.s %T/ && cp -f %S/Inputs/amdgpu_no_merge_comments-O3.s %T/
# RUN: cp -f %S/Inputs/amdgpu_no_merge_comments.ll %t.ll && %update_llc_test_checks --llc-binary cat %t.ll
# RUN: diff -u %S/Inputs/amdgpu_no_merge_comments.ll.expected %t.ll