From: Jon K. <jk...@cs...> - 2006-05-12 02:47:39
|
Hi, First, my apologies for not following up on my previous postings. I had a similar problem with rflex taking up a lot of CPU. After some discussion here and a day or two of digging through the code, I found that inserting a single usleep() in the proper place dropped by CPU usage from ~80% to 0%. Since then I got sidetracked and didn't get to submitting a patch. I'll email the details when I get into the lab tomorrow. On a related note, I found last week that Player 2.0.1 wasn't able to send commands to my rflex robot, although reading data was fine (I hadn't noticed this earlier as I hadn't tried to control the robot with 2.0.1, just read sonar and power data). I put in another one line fix and it seems to work fine. I'll send in the details of that tomorrow too. Finally, a big thanks to everyone involved in developing player/stage, my life would be more difficult without it, and you guys probably don't get enough credit for the free support and development you do! -jon On Thu, 11 May 2006, Toby Collett wrote: > Im probably repeating what was in that discussion, but basically here is > what I think has changed... > [snip] -- Jon Klippenstein, M.Sc. Candidate Robotics Research Group Dept. of Computing Science, University of Alberta |