[PATCH] D70367: Fix for AMDGPU MUL_I24 known bits calculation

Eugene Kuznetsov via Phabricator via llvm-commits llvm-commits at lists.llvm.org
Mon Dec 2 13:12:13 PST 2019


ekuznetsov139 marked an inline comment as done.
ekuznetsov139 added inline comments.


================
Comment at: llvm/test/CodeGen/AMDGPU/amdgpu-mul24-knownbits.ll:4
+; RUN: llc -mtriple amdgcn-amd-amdhsa -mcpu=gfx900 -mattr=-code-object-v3 -O2 -amdgpu-function-calls=0 < %s | FileCheck --check-prefix=GCN %s
+; GCN-NOT: -128
+; Function Attrs: alwaysinline convergent norecurse nounwind
----------------
arsenm wrote:
> ekuznetsov139 wrote:
> > arsenm wrote:
> > > Should use positive checks. I don’t know what this would exclude since we won’t emit anything with dashes
> > This would exclude
> > 
> >         global_store_dword v[0:1], v2, off offset:-128
> > 
> >           
> Ok, that’s not obvious. Positive checks are much less error prone 
This is a negative test. We are testing to make sure that the optimizer does not assume %v1 to be always equal to -32. A negative check fits right in. A positive check would be harder to write since multiple possible correct codes could be generated.



CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D70367/new/

https://reviews.llvm.org/D70367





More information about the llvm-commits mailing list