[PATCH] D107817: [RISCV][test] Add new tests for mul optimization in the zba extension with SH*ADD
Ben Shi via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Tue Aug 10 18:46:57 PDT 2021
benshi001 marked an inline comment as done.
benshi001 added inline comments.
================
Comment at: llvm/test/CodeGen/RISCV/rv64zba.ll:7
; RUN: llc -mtriple=riscv64 -mattr=+m,+experimental-zba -verify-machineinstrs < %s \
; RUN: | FileCheck %s -check-prefix=RV64IBA
----------------
jrtc27 wrote:
> craig.topper wrote:
> > craig.topper wrote:
> > > benshi001 wrote:
> > > > jrtc27 wrote:
> > > > > Ugh this is confusing, should be RV64IBZBA, not RV64IBA (which would mean integer, bitmanip and atomics, albeit in a non-canonical order, which makes no sense for this test)
> > > > Sure. I will correct that before 'git push'
> > > Why RV64IBZBA and not RV64IZBA? Should the M be mentioned for the +m?
> > >
> > > All of the B extension tests are like this and have been since they were created.
> > Please don't. All the tests should be corrected at once and not part of this patch.
> Oh, yes, RV64IZBA is what I mean here. The M probably should be there too, I think in general we either do something approximating the full arch string or we do just RV64X for the specific extension being tested (i.e. you'd drop the I for the IB and I(Z)BA ones), which would also be fine by me.
OK. I will only push the part I have added.
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D107817/new/
https://reviews.llvm.org/D107817
More information about the llvm-commits
mailing list