[PATCH] D82492: [AArch64][SVE] Guard perm and select bfloat16 intrinsic patterns
Cullen Rhodes via Phabricator via llvm-commits
llvm-commits at lists.llvm.org
Thu Jun 25 01:57:03 PDT 2020
c-rhodes added inline comments.
================
Comment at: llvm/test/CodeGen/AArch64/sve-intrinsics-perm-select-matmul-fp64.ll:1
-; RUN: llc -mtriple=aarch64-none-linux-gnu -mattr=+sve,+f64mm -asm-verbose=0 < %s -o - | FileCheck %s
+; RUN: llc -mtriple=aarch64-none-linux-gnu -mattr=+sve,+f64mm,+bf16 -asm-verbose=0 < %s -o - | FileCheck %s
----------------
fpetrogalli wrote:
> fpetrogalli wrote:
> > Nit: given that the fact that we have added the bfloat tests in the same file as the non-bfloat tests might have contributes to overseeing the problem, I wonder if it make sense to use the same approach we have done for the ACLE tests and have the tests requiring `+bf16` to be extracted into a separate `*-bfloat` file.
> FWIW, we could keep everything in one test without adding the extra +bf16 at command line invocation and by using a function attribute attached to the bfloat functions, as I have done in https://reviews.llvm.org/D82429
> FWIW, we could keep everything in one test without adding the extra +bf16 at command line invocation and by using a function attribute attached to the bfloat functions
I like this approach, I'll go with that
Repository:
rG LLVM Github Monorepo
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D82492/new/
https://reviews.llvm.org/D82492
More information about the llvm-commits
mailing list