Don't have Telegram yet? Try it now!
https://hgpu.org/?p=7594
Accurate CUDA Performance Modeling for Sparse Matrix-Vector Multiplication