After recently upgrading to pints 0.4.1, the CMAES algorithm is no longer sped up by the set_parallel command.
Although the process is parallelised (via inspection of processes with top), the time taken for optimisation with set_parallel(True) and set_parallel(False) is the same for an equivalent number of iterations.
After recently upgrading to pints 0.4.1, the CMAES algorithm is no longer sped up by the set_parallel command.
Although the process is parallelised (via inspection of processes with top), the time taken for optimisation with set_parallel(True) and set_parallel(False) is the same for an equivalent number of iterations.