CUDA and CuBLAS GPU Matrices in Clojure

Source code: https://github.com/uncomplicate/neanderthal

Sadly, last I checked, it only had bindings for Linux which was probably the right thing to do but made it hard to try on my Mac workstation. Looks really cool though, like part of Clojure's answer to Julia, which is also pretty cool.

It has bindings for OSX, Windows, and Linux. Works fine on my Mac.

When Neanderthal was first announced, I was very much in the "why isn't it core.matrix compatible?" camp (nobody wants fragmentation, and conoatibility is a good thing!), but after watching Dragan's talk at last years EuroClojure, I changed my mind because I love what non-compatibility made possible. I've yet to actually use Neanderthal (I just don't currently have those workloads, but I do have some projects in mind), but I love what I see and it seems to be getting better and better. Thanks Dragan!

Someone in this thread might have the answer:

What is currently the best bang-for-buck CUDA card in mid-2017? 1080 Ti?

I need an upgrade for Premiere, which has become increasingly demanding on my GTX 960 in the last few CC updates.

Several years ago I took CUDA course on Udacity and I thought how interesting it would be to write the same exercises in Lisp. Now this is possible, great!