SciML / DiffEqGPU.jl

GPU-acceleration routines for DifferentialEquations.jl and the broader SciML scientific machine learning ecosystem
https://docs.sciml.ai/DiffEqGPU/stable/
MIT License
284 stars 29 forks source link

Python support for DiffEqGPU #273

Open sibyjackgrove opened 1 year ago

sibyjackgrove commented 1 year ago

Is there any plan to add Python support for DiffEqGPU.jlsimilar to how DifferentialEquations.jl can be used from Python using diffeqpy?

utkarsh530 commented 1 year ago

I haven't tried it yet. Maybe post an MWE? Something like this should work: https://scicomp.stackexchange.com/questions/36994/cuda-python-for-numerical-integration-and-solving-differential-equations

ChrisRackauckas commented 1 year ago

I think they mean like what we have with R. https://www.stochasticlifestyle.com/gpu-accelerated-ode-solving-in-r-with-julia-the-language-of-libraries/ . The key thing we are missing there is that MTK tracing doesn't work in Python.

sibyjackgrove commented 1 year ago

I think they mean like what we have with R. https://www.stochasticlifestyle.com/gpu-accelerated-ode-solving-in-r-with-julia-the-language-of-libraries/ . The key thing we are missing there is that MTK tracing doesn't work in Python.

Yes, this is what I was referring to. It's unfortunate that it is not supported in Python.

ChrisRackauckas commented 1 year ago

It has to do with the wrapper library, not necessarily the GPU library. https://github.com/SciML/diffeqpy/issues/78 and https://github.com/SciML/diffeqpy/issues/57