Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

GCC Configuration #350

Open
spophale opened this issue Jun 16, 2021 · 0 comments
Open

GCC Configuration #350

spophale opened this issue Jun 16, 2021 · 0 comments
Assignees
Labels
enhancement New feature or request

Comments

@spophale
Copy link
Collaborator

I don't know whether you also want to do this separation also for GCC; by default, GCC compiles for all configured offloading devices - which is usually "none" or a specific one, but the compilers by SUSE and Debian/Ubuntu provide (if the necessary optional packages are installed) support for both AMD GCN and nvptx. – The binaries are multi-arch binaries, which support all compiled-in offloading targets (+ host fallback).

GCC

nvptx: -foffload=nvptx-none can be used to force nvptx offloading (and disabling other offloading targets); GCC (since GCC 11?) often requires -foffload=-latomic (or -foffload=nvptx-none=-latomic) to link libatomic.
AMD GCN: -foffload=amdgcn-amdhsa + -foffload=-march=fiji or likewise for gfx900, gfx906, or gfx908 (or -foffload=amdgcn-amdhsa=-march=... - if you compile for multiple offload targets at the same time.)

@spophale spophale added the enhancement New feature or request label Jun 16, 2021
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
enhancement New feature or request
Projects
None yet
Development

No branches or pull requests

2 participants