Landing: ea6244c6f8db

← All landings

Project / Subsystem

gcc / tree-ssa-math-opts

Date

2026-07-09

Author

Philipp Tomsich

Commit

ea6244c6f8db7663aca4e684c3105325d8282ede

Source

github

Perf win

Yes

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: tree-ssa-math-opts
patch_id: —
commit_hash: ea6244c6f8db7663aca4e684c3105325d8282ede
source_type: github
headline: GCC separates FMA deferral control for optimizations
tldr: GCC introduces a new parameter to control FMA deferral independently from reassoc reordering, benefiting targets like AArch64.
author: Philipp Tomsich
outcome: committed
performance_win: true
breaking_change: false
series_id: —
series_parts: []
tags: • optimization
• aarch64
• floating-point
• fma
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/ea6244c6f8db7663aca4e684c3105325d8282ede
date: 2026-07-09T00:00:00.000Z

GCC’s tree-ssa-math-opts pass now allows finer control over floating-point fused multiply-add (FMA) optimizations. A new parameter, widening-mul-defer-fma, is introduced to gate the deferral of multiply-accumulate operations, separate from the avoid-fma-max-bits parameter which controls reassoc reordering. This change prevents performance regressions on targets like AArch64, where disabling reassoc reordering to avoid FMA also unnecessarily disabled other beneficial optimizations.