Hi,

I am running Gazebo 11 (with DART 6) on a cluster with 32 CPUS and 32 GB RAM and it runs at approximately the same speed as on my 6-CPU laptop. When I run top, I see that only a tiny bit of the resources is used. My realtimeupdate_rate is set to 0 such that the simulation will run as fast as it can (as specified here).

I guess this issue is because things aren't parallelized. I found that the ODE offers parallelization here, however I'd rather stick with the DART engine because it has proven to work better for my research (in robotic grasping). Will I need to switch my physics engine back to ODE to make full use of these resources?

Any comments are much appreciated. :-)