Add GPU execution support?

[v1gnesh]

Would it be possible to add GPU execution support?
GPU's are well known for parallellism & working with the cuDF / RAPIDS ecosystem, is it possible to add GPU parallel processing capability in Rayon?

[cuviper]

GPU parallelism is powerful, but not a good match for rayon's work-stealing model. I think you'd need something designed for homogenous workloads, and it probably needs compiler support to mix different target machines, if you want the host CPU to transparently "call" functions on the GPU.

[v1gnesh]

Ok ... do you know of any crates currently working on GPU parallellism, or core-agnostic scheduling?
Also, does the Rust compiler support execution on GPUs?

[cuviper]

do you know of any crates currently working on GPU parallellism,

There are a few crates that bind OpenCL or CUDA, but I haven't used them.

or core-agnostic scheduling?

Not that I know of.

Also, does the Rust compiler support execution on GPUs?

There's an unstable target for nvptx64-nvidia-cuda, but it doesn't seem like anyone is working on it.
Here's the tracking bug: rust-lang/rust#38789

[Walther]

Related and probably of interest: https://github.com/EmbarkStudios/rust-gpu