Landing: 5c556f9f539a

← All landings

Project / Subsystem

gcc / i386

Date

2026-05-26

Author

liuhongt

Commit

5c556f9f539a9493fea9401f2e106bfd8be9ed5e

Source

github

Perf win

Yes

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: i386
patch_id: —
commit_hash: 5c556f9f539a9493fea9401f2e106bfd8be9ed5e
source_type: github
headline: i386: Disable gather optimizations for Diamond Rapids
tldr: Gather optimizations are now disabled for Diamond Rapids processors to improve pipeline utilization.
author: liuhongt
outcome: committed
performance_win: true
breaking_change: false
series_id: —
series_parts: []
tags: • i386
• optimization
• diamond rapids
• gather
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/5c556f9f539a9493fea9401f2e106bfd8be9ed5e
date: 2026-05-26T00:00:00.000Z

On Diamond Rapids (DMR) architecture, gather emulation achieves optimal pipeline utilization and parallelism with 2/4-element vectors This commit disables the use_gather_2parts and use_gather_4parts optimizations for the Diamond Rapids architecture. This adjustment aims to improve performance on DMR by using more efficient gather implementations.