Summary
-------

This distribution contains code to reproduce the results in: 

[1] M. Filippone and R. Engler. Enabling scalable stochastic gradient-based inference for Gaussian processes by employing the Unbiased LInear System SolvEr (ULISSE). January 2015. arXiv:1501.05427.


More details
-------

The folder fast_cmvps/ contains an adaptation of the code in http://www.cs.ubc.ca/~awll/nbody_methods.html to run the conjugate gradient algorithm with fast kernel matrix-vector procucts based on 'N-body' methods.
In this folder you need to compile wrapped.c using the R CMD SHLIB command.

In the other folder there are scripts to compile the .c files. Flags in the .r files determine whether covariance matrix-vector products should be done on GPUs or CPUs. The scripts might need tweaking depending on the hardware.