Don't have Telegram yet? Try it now!
https://hgpu.org/?p=16914
An N log N Parallel Fast Direct Solver for Kernel Matrices