it is a boosting method for kernel learning which is easy to parallelize and very fast. Not available yet. I will let you know.