Don't have Telegram yet? Try it now!
https://hgpu.org/?p=6294
CUDA and OpenCL-based asynchronous PSO