They should compare with other multithreading and GPU approaches for SAT/SMT solving (like https://www.win.tue.nl/~awijs/articles/parafrost_gpu.pdf https://www.

They should compare with other multithreading and GPU approaches for SAT/SMT solving (like https://www.win.tue.nl/~awijs/articles/parafrost_gpu.pdf https://www.win.tue.nl/~awijs/articles/parafrost_gpu.pdf from Armin Biere, or other works from Mate Soos). There has been a lot of research in this direction. Other old HN thread (2017) with relevant comments https://news.ycombinator.com/item?id=13667380 https://news.ycombinator.com/item?id=13667380 from actual experts.