ROS Resources: Documentation | Support | Discussion Forum | Index | Service Status | ros @ Robotics Stack Exchange
Ask Your Question

SMACH slows down after hours of execution of large FSM on PR2

asked 2011-07-25 01:37:28 -0500

mbollini gravatar image

updated 2014-11-22 17:05:38 -0500

ngrennan gravatar image

When I run my SMACH code on the PR2 for a long time, about 90 minutes or more, it slows down considerably. Everything is in one giant (albeit nested) SMACH fsm with about 125 states. The slowdown affects everything in the (python) process, including state transitions and all other bits of code. When running top I don't see any big CPU or memory usage increase.

Have you guys run into this on your SMACH codes? I've been bumping into this more and more as the code crashes less often (meaning longer code runtimes). Restarting the robot (both hardware and software) does not seem to prevent this from happening again.

This happens even when smach_viewer is not running.

edit retag flag offensive close merge delete

1 Answer

Sort by ยป oldest newest most voted

answered 2011-07-25 06:02:11 -0500

Wim gravatar image

updated 2011-07-25 06:16:12 -0500

dornhege gravatar image

We've been running smach state machines with a few dozen states for days-weeks at a time, and never saw this behavior. A few things you could check:

  • Verify that it is actually smach that is getting slower (e.g. using print's), and not the ros communication in/out of smach
  • Check the overall system load when this is happening (using for example top)
  • Does your smach process take up a full core?
edit flag offensive delete link more

Question Tools

1 follower


Asked: 2011-07-25 01:37:28 -0500

Seen: 494 times

Last updated: Jul 25 '11