How to run a single simulation of ROS on MultiCores?
Hi, I am working with Reinforcement Learning train a UR5 to reach a target position. I am using a 64 Core Processor which brings the "Real Time Factor" of the simulation >2 and a "frame rate of 62" which makes the simulation X2 times faster than the real time! but what I notice is when training is going on the simulation and my code only occupies 20% of the CPU which I believe is inefficient. Moreover, the main problem is when a other person tries to run his code on the same server the "Real Time Factor" goes down to 1. I beleive that this is because the ROS simulation is running on just 1 CPU. Is there a way to run this one simulation on all the cores on the CPU to get a higher Real Time Factor and a higher FPS?
add a comment
