Landing: e870be66c6ad

← All landings

Project / Subsystem

gcc / i386

Date

2026-07-09

Author

liuhongt

Commit

e870be66c6adb856b2ee8869801f8c4542b3a914

Source

github

Perf win

Yes

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: i386
patch_id: —
commit_hash: e870be66c6adb856b2ee8869801f8c4542b3a914
source_type: github
headline: i386: Add vector lrint expanders for conversions
tldr: Enables vectorizer to use new conversion instructions for 'lrint' operations.
author: liuhongt
outcome: committed
performance_win: true
breaking_change: false
series_id: —
series_parts: []
tags: • i386
• vectorization
• optimization
• intrinsics
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/e870be66c6adb856b2ee8869801f8c4542b3a914
date: 2026-07-09T00:00:00.000Z

This commit exposes existing vector conversion instructions, vcvtps2qq and vcvtpd2dq, through lrint optabs. This allows the vectorizer to utilize these specialized instructions instead of falling back to less efficient scalar conversion sequences when determining the nearest integral value for floating-point numbers. New tests have been added to verify this functionality.