[llvm] [AMDGPU][GlobalISel] Enable vector reductions (PR #131413)

[Matt Arsenault via llvm-commits]

================
@@ -0,0 +1,3761 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 5
----------------
arsenm wrote:

By main codegen directory, I mean llvm/test/CodeGen/AMDGPU/vector_reduce_<op>.ll, not this new subdirectory convention 

https://github.com/llvm/llvm-project/pull/131413