Don't have Telegram yet? Try it now!
https://hgpu.org/?p=7116
The Boat Hull Model: Adapting the Roofline Model to Enable Performance Prediction for Parallel Computing