HomePhabricator

Change semantics of fadd/fmul vector reductions.

Authored by sdesmalen on Jun 11 2019, 1:22 AM.

Description

Change semantics of fadd/fmul vector reductions.

This patch changes how LLVM handles the accumulator/start value
in the reduction, by never ignoring it regardless of the presence of
fast-math flags on callsites. This change introduces the following
new intrinsics to replace the existing ones:

llvm.experimental.vector.reduce.fadd -> llvm.experimental.vector.reduce.v2.fadd
llvm.experimental.vector.reduce.fmul -> llvm.experimental.vector.reduce.v2.fmul

and adds functionality to auto-upgrade existing LLVM IR and bitcode.

Reviewers: RKSimon, greened, dmgreen, nikic, simoll, aemerson

Reviewed By: nikic

Differential Revision: https://reviews.llvm.org/D60261

llvm-svn: 363035

Details

Committed
sdesmalenJun 11 2019, 1:22 AM
Reviewer
nikic
Differential Revision
D60261: Change semantics of fadd/fmul vector reductions.
Parents
rGe2acbeb94cf2: [clangd] Fix gcc warning by removing extra ";"
Branches
Unknown
Tags
Unknown