Re: physics

well, as i said, I've tried many kinds of settings, including 32ms step simulation. the "run real time" option was always enabled. I guess I could write my own 5s limit using the system time. using 32ms steps actually seemed to be totally different to the contest-simulation and caused far more crashes, a reason why I switched to 16ms steps. but I'll stick to your recomendation and hope that i get more accurate results while testing. Is it possible that the simulation varies from computer to computer, depending on processor speed and graphic-card power?
  ----- Original Message ----- 
  From: Olivier Michel 
  To: [email protected] 
  Sent: Wednesday, November 26, 2003 6:38 PM
  Subject: Re: [webots-contest] physics


  Hi,

  I would recommand you to use a basic simulation step of 32 ms (as this 
  is the value used for the contest matches). You may experience different 
  behaviors in the contest rounds because the robot controllers run 
  asynchronously with the simulator. Hence I would recommand you to check 
  the "run real time" option as well, so that you will be as close as 
  possible to the contest conditions (this is especially useful to take 
  into account the 5 seconds limit).

  Using these options, you local simulation should run almost the same as 
  on the contest server. Providing a "contest mode" wouldn't help much I 
  think...