Have you thought of making a portable header version of this for use in c++/c for general purpose use outside of R?
Charles Determan
@cdeterman
Hi @jaredhuling , that is actually what ViennaCL backend is for (which I have conveniently package in the RViennaCL package for R users). The main website for ViennaCL is here. I am simply using it to allow convenient GPU use in R.
Jared Huling
@jaredhuling
@cdeterman thanks for the pointer! I was unaware of this but it is exactly what I had in mind.
Francisco Bischoff
@franzbischoff
Salutations
Dummy question for an GPU-starter-guy: does matrix calculations using GPU is supposed to be faster than CPU right? Because I've been testing and seeing posts showing GPU slower than CPU. The trick for GPU would be using hundreds of threads?
Charles Determan
@cdeterman
@franzbischoff
@franzbischoff
Silly app, sorry for posts. I would like to see the posts you are referring. Also, feel free to submit an issue on github. There are many moving parts that could explain the speed discrepancy. @franzbischoff
Francisco Bischoff
@franzbischoff
sure, when I get back to GPU implementation of the algorithm I'm working in :) thanks for the answer @cdeterman