commit: 02d3738be92eac38cebfb7b670673abb9538ca76
[log]
author: Florian Hahn <flo@fhahn.com>
Tue Jul 15 16:19:27 2025 +0200
committer: GitHub <noreply@github.com>
Tue Jul 15 15:19:27 2025 +0100
tree: d4a2c70412296081241b452fdd93c34737a05210
parent: 9f7937401b62d95f2ebe86072d6f277eb9745cd0 [diff]

[AArch64,TTI] Remove RealUse check for vector insert/extract costs. (#146526)

getVectorInstrCostHelper would return costs of zero for vector
inserts/extracts that move data between GPR and vector registers, if
there was no 'real' use, i.e. there was no corresponding existing
instruction.

This meant that passes like LoopVectorize and SLPVectorizer, which
likely are the main users of the interface, would understimate the cost
of insert/extracts that move data between GPR and vector registers,
which has non-trivial costs.

The patch removes the special case and only returns costs of zero for
lane 0 if it there is no need to transfer between integer and vector
registers.

This impacts a number of SLP test, and most of them look like general
improvements.I think the change should make things more accurate for any
AArch64 target, but if not it could also just be Apple CPU specific.

I am seeing +2% end-to-end improvements on SLP-heavy workloads.

PR: https://github.com/llvm/llvm-project/pull/146526

llvm/lib/Target/AArch64/AArch64TargetTransformInfo.cpp[diff]
llvm/lib/Target/AArch64/AArch64TargetTransformInfo.h[diff]
llvm/test/Analysis/CostModel/AArch64/reduce-and.ll[diff]
llvm/test/Analysis/CostModel/AArch64/reduce-or.ll[diff]
llvm/test/Analysis/CostModel/AArch64/reduce-xor.ll[diff]
llvm/test/Analysis/CostModel/AArch64/shuffle-other.ll[diff]
llvm/test/Analysis/CostModel/AArch64/shuffle-select.ll[diff]
llvm/test/Analysis/CostModel/AArch64/sve-intrinsics.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/ext-trunc.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/external-use-icmp.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/extract-subvector-long-input.ll[Renamed from llvm/test/Transforms/SLPVectorizer/extract-subvector-long-input.ll - diff]
llvm/test/Transforms/SLPVectorizer/AArch64/extractelements-to-shuffle.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/gather-buildvector-with-minbitwidth-user.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/getelementptr.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/getelementptr2.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/memory-runtime-checks.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/multiple_reduction.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/phi-node-bitwidt-op-not.ll[Renamed from llvm/test/Transforms/SLPVectorizer/phi-node-bitwidt-op-not.ll - diff]
llvm/test/Transforms/SLPVectorizer/AArch64/splat-loads.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/unreachable-blocks-with-phis.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/vector-getelementptr.ll[diff]
llvm/test/Transforms/SLPVectorizer/AArch64/vectorizable-selects-uniform-cmps.ll[diff]
llvm/test/Transforms/SLPVectorizer/X86/extract-subvector-long-input.ll[Copied from llvm/test/Transforms/SLPVectorizer/extract-subvector-long-input.ll - diff]
llvm/test/Transforms/SLPVectorizer/X86/icmp-altopcode-after-reordering.ll[Renamed from llvm/test/Transforms/SLPVectorizer/icmp-altopcode-after-reordering.ll - diff]
llvm/test/Transforms/SLPVectorizer/X86/phi-node-bitwidt-op-not.ll[Copied from llvm/test/Transforms/SLPVectorizer/phi-node-bitwidt-op-not.ll - diff]
llvm/test/Transforms/SLPVectorizer/alternate-opcode-sindle-bv.ll[diff]
llvm/test/Transforms/VectorCombine/AArch64/load-extractelement-scalarization.ll[diff]

27 files changed

tree: d4a2c70412296081241b452fdd93c34737a05210

README.md

The LLVM Compiler Infrastructure

Welcome to the LLVM project!

This repository contains the source code for LLVM, a toolkit for the construction of highly optimized compilers, optimizers, and run-time environments.

The LLVM project has multiple components. The core of the project is itself called “LLVM”. This contains all of the tools, libraries, and header files needed to process intermediate representations and convert them into object files. Tools include an assembler, disassembler, bitcode analyzer, and bitcode optimizer.

C-like languages use the Clang frontend. This component compiles C, C++, Objective-C, and Objective-C++ code into LLVM bitcode -- and from there into object files, using LLVM.

Other components include: the libc++ C++ standard library, the LLD linker, and more.

Getting the Source Code and Building LLVM

Consult the Getting Started with LLVM page for information on building and running LLVM.

For information on how to contribute to the LLVM project, please take a look at the Contributing to LLVM guide.

Getting in touch

Join the LLVM Discourse forums, Discord chat, LLVM Office Hours or Regular sync-ups.

The LLVM project has adopted a code of conduct for participants to all modes of communication within the project.