At the moment the search procedure only uses a single CPU thread; it would be nice to speed things up by employing parallelism.
One obvious way to do this is to simply split up the search space, and explore each subspace on a different thread. The only difficulty is that different subspaces might take different amounts of time to explore, which can lead to an unbalanced workload.
At the moment the search procedure only uses a single CPU thread; it would be nice to speed things up by employing parallelism.
One obvious way to do this is to simply split up the search space, and explore each subspace on a different thread. The only difficulty is that different subspaces might take different amounts of time to explore, which can lead to an unbalanced workload.