How to use all processor cores in optimization?

[sambitn...@gmail.com]

I am running a multi-objective optimization using NSGA-II from Inspyred package. However, I need to run it real-time through a app front-end. The overall optimization is taking ~2 minutes which is very long for an app. Is there a way the operations can run using all processor cores and hence make it faster. I noticed it is not utilizing all the processors and just working through on one. Maybe this is a very naive question. But would love to get some perspective on this.

[Aaron Garrett]

You can use a parallel evaluator (like https://pythonhosted.org/inspyred/reference.html#inspyred.ec.evaluators.parallel_evaluation_mp) that will use all of the processors/cores that you specify. But depending on the problem, getting a solution can take a while. If you want it to be in real-time, then you might want to explore other ideas. For instance, you could have it run for a smaller population and number of evaluations, give an answer pretty quickly, and then let the user start from that answer to do it again if they want. However, that generally won't be as effective as having a larger population evolve for the same amount of time.

--
Aaron Garrett

On Fri, May 31, 2019 at 8:12 AM <sambitn...@gmail.com> wrote:

I am running a multi-objective optimization using NSGA-II from Inspyred package. However, I need to run it real-time through a app front-end. The overall optimization is taking ~2 minutes which is very long for an app. Is there a way the operations can run using all processor cores and hence make it faster. I noticed it is not utilizing all the processors and just working through on one. Maybe this is a very naive question. But would love to get some perspective on this.

--
You received this message because you are subscribed to the Google Groups "Inspyred" group.
To unsubscribe from this group and stop receiving emails from it, send an email to inspyred+u...@googlegroups.com.
To post to this group, send email to insp...@googlegroups.com.
Visit this group at https://groups.google.com/group/inspyred.
To view this discussion on the web visit https://groups.google.com/d/msgid/inspyred/ef4160b0-24dc-4ceb-9146-f310ab7c2f4e%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[sambitn...@gmail.com]

Thanks a lot Aaron. We are using parallel evaluation and working just fine. :)