Don't have Telegram yet? Try it now!
https://hgpu.org/?p=17025
An Efficient Multiway Mergesort for GPU Architectures