[PATCH] D80231: [llvm] [CodeGen] [X86] Fix issues with v4i1 instruction selection

Wed May 20 12:02:44 PDT 2020

aartbik added inline comments.

================
Comment at: llvm/test/CodeGen/X86/pr45995-2.ll:2
+; NOTE: Assertions have been autogenerated by utils/update_llc_test_checks.py
+; RUN: llc -O3  --x86-asm-syntax=intel -mtriple=x86_64-grtev4-linux-gnu -march=x86-64 -mcpu=skylake-avx512 -mattr=fma,avx512f < %s | FileCheck %s
+
----------------
RKSimon wrote:
> Both these run args look over constrained - llc -mtriple=x86_64-- -mattr=+avx512dq,+avx512vl might be enough?
Oops, I started the land already, but should have addressed this first. Apologizes.
Happy to relax this in a follow-up if needed. But just curious, what are the (dis)advantages of adding some extra args in a crash-before/pass-after regression test?
(but I agree that e.g. the syntax flag is a personal setting)

Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D80231/new/

https://reviews.llvm.org/D80231