Package Details: openmp-nvptx 10.0.1-1

Git Clone URL: https://aur.archlinux.org/openmp-nvptx.git (read-only, click to copy)
Package Base: openmp-nvptx
Description: LLVM OpenMP Runtime Library with NVPTX offloading targets enaled
Upstream URL: https://openmp.llvm.org/
Licenses: custom:Apache 2.0 with LLVM Exception
Conflicts: openmp
Provides: openmp
Submitter: danielotero
Maintainer: danielotero
Last Packager: danielotero
Votes: 0
Popularity: 0.000000
First Submitted: 2020-06-11 20:32
Last Updated: 2020-07-30 12:58

Required by (74)

Sources (2)

Pinned Comments

danielotero commented on 2020-06-11 20:46

This package should allow you to compile and run OpenMP offload code with clang. You'll need the cuda package installed.

The compilation syntax is:

clang -fopenmp -fopenmp-targets=nvptx64-nvidia-cuda [-Xopenmp-target -march=sm_70] <rest of the compiler arguments>

A warning will be raised because clang-10 still doesn't recognizes CUDA 10.2. You can ignore it or disable it (with -Wno-unknown-cuda-version).

clang-10: warning: Unknown CUDA version 10.2. Assuming the latest supported version 10.1 [-Wunknown-cuda-version]

Latest Comments

danielotero commented on 2020-06-11 20:46

This package should allow you to compile and run OpenMP offload code with clang. You'll need the cuda package installed.

The compilation syntax is:

clang -fopenmp -fopenmp-targets=nvptx64-nvidia-cuda [-Xopenmp-target -march=sm_70] <rest of the compiler arguments>

A warning will be raised because clang-10 still doesn't recognizes CUDA 10.2. You can ignore it or disable it (with -Wno-unknown-cuda-version).

clang-10: warning: Unknown CUDA version 10.2. Assuming the latest supported version 10.1 [-Wunknown-cuda-version]