Don't have Telegram yet? Try it now!
https://hgpu.org/?p=8663
Accuracy, Memory, and Speed Strategies in GPU-Based Finite-Element Matrix-Generation