Performance differences with float vs double for Real

Requests regarding how to set up experiments in ARGoS.
jharwell
Posts: 67
Joined: Thu Jan 25, 2018 4:12 am

Performance differences with float vs double for Real

Postby jharwell » Sat Nov 16, 2019 5:52 am

Hi Carlo,

Do you have any experience with seeing noticable performance gains when you use float instead of double for the Real primitive? Were the losses in simulation fidelity significant ? I haven't been able to find anything in the docs showing the conditions under which one vs. the other could/should be used.

Thanks!

- John

pincy
Site Admin
Posts: 632
Joined: Thu Mar 08, 2012 8:04 pm
Location: Boston, MA
Contact:

Re: Performance differences with float vs double for Real

Postby pincy » Sat Nov 16, 2019 5:55 am

In general, on 64bit machines it's better to use double for reasons of numerical stability. Float is for when ARGoS is crosscompiled on real robots.
I made ARGoS.

jharwell
Posts: 67
Joined: Thu Jan 25, 2018 4:12 am

Re: Performance differences with float vs double for Real

Postby jharwell » Mon Nov 25, 2019 6:51 pm

Got it, thanks!


Return to “How to... ?”