Gazebo-7 is slower than gazebo-2 [closed]
I run the same Baxter robot simulation in gazebo-2 with ros-indigo and gazebo-7 with ros-kinetic. But the simulation in gazebo-7 is much slower than gazebo-2. In gazebo-2 if I set the real time update rate to 10000, it can reach around 5000, but in gazebo-7 it can only reach 2000. So the simulation in gazebo-7 is inefficient. Is this normal? And how should I speed up it in gazebo-7? Changing max step size too much is not a good idea. it will influence the simulation. Many thanks.
Or it is more physically accurate and needs more calculations per time-step. With a fixed performance PC, that would lead to lower (perceived) performance.
I don't think you can draw the conclusion like you do from this single observation.
In any case, I believe this is a Gazebo-specific question, and those should be asked on a more appropriate forum.
If you do post your question over at
answers.gazebosim.org
, please post a comment with a link to it here on ROS Answers, so we keep things connected.I put it on gazebo answer http://answers.gazebosim.org/question...
@gvdhoorn I also think maybe more accurate is the reason. But it is not good for many applications. If the simulation is in the same speed or slower than reality, accuracy is meaningless for many applications. I can run it directly in reality, which should be the most accurate.
Well, that depends.
I for one am very happy that I can use a simulated robot and do anything with it, without running the risk of breaking the real one. Simulated robots also give you ..
.. multiple copies of the same robot hw for free.
I don't think simulations that are at or below real-time are meaningless. Perhaps for your application(s), but such a general statement can (again) not be made like this.
@gvdhoorn I agree. So I said meaningless for many applications.