HigherOrderCO
FFI for rustlang like a tokio runtime function where i can offload the specific function code to gpu
Is your feature request related to a problem? Please describe. since it is written in rustlang, possible to provide a way to use bend for performance sensitive / parallelizable stuff to run in cuda etc?
Describe the solution you'd like some code i can write in either rust or bend and have it parallelize for me in either cpu or gpu as i specified it as a function wrapped like tokio.
Describe alternatives you've considered none
Additional context none
The existing runtimes are in C and Cuda. The rust runtime is only a reference and isn't parallel.
Regardless, yes it's possible and we want to have some bindings in other languages in the future, although we don't have any concrete plans yet.
