Cuda accelerated version of Alternating Direction Method of Multipliers Lasso

Used a combination of CUBLAS, some open source code, and my own kernels to create this project;

https://github.com/OlegKonings/GPU_ADMM_Lasso

still in the beta phase, but I thought some on here might enjoy it. I have found this forum very helpful.

Obviously the GPU version is better optimized than the CPU version (which I also wrote from scratch as I did not have access to any Linear Algebra libraries).

It was a lot of fun, and I will be adding more to the project in the next few weeks.