[llvm] [VectorCombine][AMDGPU] Shrink demanded vector loads (PR #202501)

[Krzysztof Drewniak via llvm-commits]

================
@@ -0,0 +1,206 @@
+; RUN: opt < %s -passes=vector-combine -S -mtriple=amdgcn-amd-amdpal -mcpu=gfx1201 | FileCheck %s
----------------
krzysz00 wrote:

Nit: autogenerate your tests here (and possibly doe pre-commit)

https://github.com/llvm/llvm-project/pull/202501