Don't have Telegram yet? Try it now!
https://hgpu.org/?p=22164
Automated Partitioning of Data-Parallel Kernels using Polyhedral Compilation