Skip to content

Restore CUDA.jl v5 support #449

@nicholaskl97

Description

@nicholaskl97

Is your feature request related to a problem? Please describe.

As described in SciML/NeuralLyapunov.jl#175, AdvancedHMC doesn't yet support CUDA v6, so removing support for CUDA v5 means DiffEqGPU cannot be used with it, which was causing problems for NeuralLyapunov (probably less so now that BPINNs have been moved to an extension in NeuralPDE v6).

Describe the solution you’d like

Since the CUDA v6 claims there should be ideally no breaking changes from v5, removing support for v5 in #441 was probably unnecessarily conservative.

Describe alternatives you’ve considered

Advanced HMC could also just add CUDA v6 to their compatibility (TuringLang/AdvancedHMC.jl#493).

Additional context

@ChrisRackauckas, we spoke about this in yesterday's meeting. I'm making this issue in case you still want to go ahead with it, as it may help someone else, but as I mention above switching NeuralLyapunov to use NeuralPDE v6 should also fix my issue.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type
    No fields configured for issues without a type.

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions