Skip to content

[FEA] Clarify or support CPU fallback when CUDA is unavailable #870

@maifeeulasad

Description

@maifeeulasad

Is your feature request related to a problem? Please describe.
When using @cuda.jit, execution requires a CUDA-capable device at runtime. In environments where CUDA is unavailable (e.g., CPU-only systems, CI pipelines, or heterogeneous deployments), the code fails instead of degrading gracefully. This forces developers to maintain separate CPU (@njit) and GPU (@cuda.jit) implementations, increasing complexity and duplication.

Describe the solution you'd like
Provide an optional CPU fallback mechanism (in the decorator itself, @cuda.jit(fallback_to_cpu=True)) which defaults to false. So the exisiting behaviour doesn't change. But as soon as we enable it. Voila! We are good.

Describe alternatives you've considered
n/a

Additional context
Willing to work on this one!

Metadata

Metadata

Assignees

No one assigned

    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