Don't have Telegram yet? Try it now!
https://hgpu.org/?p=14627
Optimizing GPU-accelerated Group-By and Aggregation