ES optimizers have the reputation of being slow, so libcmaes tries to yield the best performances both in terms of computational cost of the optimization step, and wrt the number of function evaluations.

This ticket is about benchmarking libcmaes against other existing implementations. My current list contains:

• Shark (http://image.diku.dk/shark/sphinx_pages/build/html/rest_sources/tutorials/algorithms/cma.html)
• C and Python from https://www.lri.fr/~hansen/cmaes_inmatlab.html

It will be possible to add others as needed.