Landing: 0a1f6acb90e2

← All landings

Project / Subsystem

gcc / amdgcn

Date

2026-09-24

Author

Andrew Stubbs

Commit

0a1f6acb90e27ac0b2984ad6841186ca54bc5725

Source

github

Perf win

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: amdgcn
patch_id: —
commit_hash: 0a1f6acb90e27ac0b2984ad6841186ca54bc5725
source_type: github
headline: amdgcn: Fix 64-bit vector multiply bug
tldr: Corrects a mathematical error in the 64-bit vector multiply operation for the AMD GCN architecture.
author: Andrew Stubbs
outcome: committed
performance_win: false
breaking_change: false
series_id: —
series_parts: []
tags: • amdgcn
• gpu
• vectorization
• bugfix
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/0a1f6acb90e27ac0b2984ad6841186ca54bc5725
date: 2026-09-24T00:00:00.000Z

A bug in the compiler’s handling of 64-bit vector multiplication on AMD GCN GPUs has been fixed. The erroneous ‘hi*hi’ step in the multiplication logic has been removed, ensuring correct results for these operations.