[llvm] [NFC][SLP] Precommit tests for the phantom load saving (PR #228901)

[Dmitry Sidorov via llvm-commits]

================
@@ -0,0 +1,1536 @@
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py UTC_ARGS: --version 6
+; RUN: opt -mtriple=amdgpu9.0a-amd-amdhsa -passes=slp-vectorizer < %s | llc -mtriple=amdgpu9.0a-amd-amdhsa | FileCheck -check-prefix=GFX90A %s
+; RUN: opt -mtriple=amdgpu10.30-amd-amdhsa -passes=slp-vectorizer < %s | llc -mtriple=amdgpu10.30-amd-amdhsa | FileCheck -check-prefix=GFX1030 %s
+; RUN: opt -mtriple=amdgpu12.50-amd-amdhsa -passes=slp-vectorizer < %s | llc -mtriple=amdgpu12.50-amd-amdhsa | FileCheck -check-prefix=GFX1250 %s
----------------
MrSidims wrote:

> It is hard to maintain such tests

Only in a case if there are multiply after every patch. IMHO having a reasonable amount of such tests won't harm. IMHO if we had such tests could help to resolve [concern](https://github.com/llvm/llvm-project/pull/226117#issuecomment-5908037440) raised by @tommat01 - not every change require a performance run on an appropriate hardware if asm diff is visible in the PR and also, since LLVM bot would notify about a change in a backend test, would bring more attention for reviewers. 

https://github.com/llvm/llvm-project/pull/228901