Landing: 84d93bc97a36

← All landings

Project / Subsystem

gcc / openmp

Date

2026-09-26

Author

Tobias Burnus

Commit

84d93bc97a3610446c4e19fa84d5a10f8b668358

Source

github

Perf win

Breaking

View on GitHub →

All attributes

project: gcc
subsystem: openmp
patch_id: —
commit_hash: 84d93bc97a3610446c4e19fa84d5a10f8b668358
source_type: github
headline: OpenMP: Enable 'omp groupprivate' for nvptx offload
tldr: OpenMP's `groupprivate` clause is now supported for nvptx offload targets.
author: Tobias Burnus
outcome: committed
performance_win: false
breaking_change: false
series_id: —
series_parts: []
tags: • openmp
• nvptx
• gpu
• offload
• fortran
discussion_id_link: —
bugzilla_pr: —
url: https://github.com/gcc-mirror/gcc/commit/84d93bc97a3610446c4e19fa84d5a10f8b668358
date: 2026-09-26T00:00:00.000Z

The OpenMP groupprivate clause, which provides a copy of a variable per thread team, is now enabled for offloading to NVIDIA’s nvptx devices. This feature aligns with Nvidia’s ‘shared’ variable concept. The commit adds support in the NVPTX backend and Fortran frontend, while issuing a ‘sorry’ message for other devices like GCN or when accessing from the host unless specific conditions are met. This enables more flexible data scoping for parallel regions executed on GPUs.