Landing: 46ef72e150b7

← All landings

Project / Subsystem

gcc / i386

Date

2026-07-22

Author

Hu, Lin1

Commit

46ef72e150b7ecee62d193a5382547f3b3cfebca

Source

github

Perf win

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: i386
patch_id: —
commit_hash: 46ef72e150b7ecee62d193a5382547f3b3cfebca
source_type: github
headline: i386: Update Suffix for AVX10.2 SAT CVT scalar Intrinsics
tldr: Corrects intrinsic naming for AVX10.2 scalar SAT CVT operations to align with existing conventions.
author: Hu, Lin1
outcome: committed
performance_win: false
breaking_change: false
series_id: —
series_parts: []
tags: • i386
• vectorization
• intrinsics
• avx10
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/46ef72e150b7ecee62d193a5382547f3b3cfebca
date: 2026-07-22T00:00:00.000Z

The intrinsic names for AVX10.2 scalar SAT CVT (Saturate Convert) operations have been updated to conform to the established naming convention for scalar convert intrinsics. Previously, they used an _ep[i|u]{32,64} suffix, but the standard is now _[i|u]{32,64}. This change renames the affected intrinsics and adds aliases for signed variants with the _si{32,64} suffix, ensuring consistency with AVX-512F and later standards.