OpenCL / CUDA Software For Linux

I would like for this to be a list of linux software that uses OpenCL or CUDA and what it does and how it does that. The main reason for me is that I have GPUs that I could use for stuff in my server and I bet I am not the only one with a spare GPU lying around. I will be looking into such software over the next few days and maybe come up with something to put here but with so many linux tinkerers around here I thought why not just get it going.

I don't intend to keep updating the first post like the Linux Software List (thanks to @anon97550313) because I am a lazy mofo but it would be nice to get a quick overview, so here is a preset for you to copy and paste:

Summary
_Name of the program:_
**The Quick Brown Fox Simulator**

_Type of GPU acceleration:_
**CUDA / OpenCL / both**

_General classification of the software:_
**VR terminal emulator written in BASIC**

_Usage scenario / how-to:_
**So you lorem ipsum the damn thing and then when everything is dolor sit amet, you will be able to consetetur sadipscing elitr basically any type of sed diam nonumy, which is very eirmod tempor invidunt because banana.**

_Homepage:_
**https://theplaceholderproject.org**

Please don't start any kind of discussion or ask about your problem here.
This is supposed to be just a list.
Thank you.

The typical stuff for coding CUDA libraries is nsight which is the Nvidia IDE that is based on eclipse for linux.

MATLAB also has very easy and good integration to CUDA. More high level thus less well performing that developing on you own but very easy to use if you know MATLAB.

Then there is a whole list that use OpenCL somehow: https://en.wikipedia.org/wiki/List_of_OpenCL_applications

And all 3 major GPU vendors have an SDKs to use OpenCL.

torch, caffe

1 Like

If your not a MATLAB person, or don't have the money for a license, Python has its own high level implementations of CUDA and OpenCL. I have never used them so I am not sure how much control they give you compared to the source language.

PyCUDA: https://documen.tician.de/pycuda/
PyOpenCL: https://documen.tician.de/pyopencl/