Deprecated, please use AMDGPU.jl
Support for compiling and executing native Julia kernels on AMD GPUs.
| Documentation | Build Status |
|---|---|
 |
  |
The package can be installed with the Julia package manager.
From the Julia REPL, type ] to enter the Pkg REPL mode and run:
pkg> add AMDGPUnative
Or, equivalently, via the Pkg API:
julia> import Pkg; Pkg.add("AMDGPUnative")The package is tested against, and being developed for, Julia 1.3 and above.
Only 64-bit Linux is supported and working at this time, until ROCm is ported
to other platforms. It is recommended to use a version of Julia with LLVM 9.0
or higher. This package is under active maintenance and is reasonably
complete, however not all features (and especially performance) are up to par
with CUDAnative.
| Feature | Supported | Notes |
|---|---|---|
| Host-side kernel launches | ✔️ | See #58 |
| Dynamic parallelism | ❌ | |
| Local (shared) memory | ❌ | |
| Coarse-grained memory | ❌ | |
| Page-locked (pinned) memory | ❌ |
Usage questions can be posted on the Julia Discourse forum under the GPU domain and/or in the #gpu channel of the Julia Slack.
Contributions are very welcome, as are feature requests and suggestions. Please open an issue if you encounter any problems.
AMDGPUnative would not have been possible without the work by Tim Besard and contributors to CUDAnative.jl and LLVM.jl.
AMDGPUnative.jl is licensed under the MIT License.