projects / platform / upstream / llvm.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 2d33327) | patch
[LoopVectorize][CostModel] Update cost model for fmuladd intrinsic
authorRosie Sumpter <rosie.sumpter@arm.com>
Tue, 12 Oct 2021 08:51:42 +0000 (09:51 +0100)
committerRosie Sumpter <rosie.sumpter@arm.com>
Wed, 24 Nov 2021 08:50:05 +0000 (08:50 +0000)
commitdf32a39dd0f68383a1685a4571715edb70664969
tree4953c1727f5c4087e3b659dd4a1106a0a3ec18aftree | snapshot
parent2d33327f9d4c154e7454c2d830c1caa8e6850f4fcommit | diff
[LoopVectorize][CostModel] Update cost model for fmuladd intrinsic

This patch updates the cost model for ordered reductions so that a call
to the llvm.fmuladd intrinsic is modelled as a normal fmul instruction
plus the cost of an ordered fadd reduction.

Differential Revision: https://reviews.llvm.org/D111630
llvm/lib/Transforms/Vectorize/LoopVectorize.cpp diff | blob | history
llvm/test/Transforms/LoopVectorize/AArch64/strict-fadd-cost.ll diff | blob | history
Domain: System / Toolchain;