JuliaGPU / JuliaGPU/KernelAbstractions.jl
Add API for device selection
- Dominant language
- Julia
- Stars
- 523
- Forks
- 88
- Avg merge
- 1d 11h
- Merged PRs (30d)
- 25
Description
From a short conversation with @vchuravy I learned that there is currently no way of selecting between different GPUs by the same vendor on a single host using KernelAbstractions. It seems that KA.jl always defaults to the GPU with the lowest index.
CUDA.jl offers this feature already, see https://cuda.juliagpu.org/stable/usage/multigpu/, I'm not sure about AMDGPU.jl and oneAPI.jl. KA.jl should maybe offer some backend agnostic API for this.
Contributor guide
No contributing guide indexed for this repository
Research direction
Read the linked CUDA.jl multigpu documentation and inspect KernelAbstractions.jl's current backend and device-selection behavior. Compare the selection capabilities of AMDGPU.jl and oneAPI.jl, then determine what a backend-agnostic API would need to support selecting among same-vendor GPUs and how that behavior should be tested.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- julia
- Domain
- backend-api-design, hpc
- Issue type
- Feature
- Difficulty
- 5/5
- Estimated time
- Over a week
- Activity status
- Stale
- Clarity
- Needs clarification
- Newbie friendliness
- 25/100