Landing: 398bdccce49e

← All landings

Project / Subsystem

gcc / gcc/ira

Date

2026-06-12

Author

Stefan Schulze Frielinghaus

Commit

398bdccce49e0ed5774cf05f7f1fa5aa00f54afe

Source

github

Perf win

Yes

Breaking

No

View on GitHub →

All attributes

project: gcc
subsystem: gcc/ira
patch_id: —
commit_hash: 398bdccce49e0ed5774cf05f7f1fa5aa00f54afe
source_type: github
headline: IRA aligns equivalence rejection with LRA policies
tldr: The IRA pass in GCC now anticipates LRA's rejection of certain register equivalences, reducing unnecessary spilling and improving code quality for PIC.
author: Stefan Schulze Frielinghaus
outcome: committed
performance_win: true
breaking_change: false
series_id: —
series_parts: []
tags: • ira
• lra
• optimization
• pic
• s390
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/398bdccce49e0ed5774cf05f7f1fa5aa00f54afe
date: 2026-06-12T00:00:00.000Z

The Instruction Register Allocator (IRA) in GCC now harmonizes its handling of register equivalences with the Local Register Allocator (LRA). Previously, IRA might unnecessarily spill pseudo-registers with equivalences that LRA would ultimately reject, especially in Position-Independent Code (PIC) scenarios. By aligning IRA’s costing function to reject these equivalences earlier, the compiler avoids generating inefficient code that creates stack frames solely for these spills, resulting in smaller and faster binaries.