erin.taylor@jhuapl.edu a écrit : Hi Erin,
Thank you for the release of 6.1. I dropped it in yesterday to replace 6.0, but to my dismay I found that it increased my application run time by a factor of about 3 without changing any of my code. This is what I am doing: I am calling orekit commands from within a matlab script to numerically propagate elements input in EME2000 with atmospheric drag, solar radiation, earth, sun, and moon gravity. I propagate in slave mode, controlling the propagator with a loop in Matlab, for 30 minutes at 1 minute steps, then output the coordinates at each time step in the ITRF2008 frame.
One thing that may be involved in your case is the ITRF frame. Could you check that you use a frame retrieved from one of the factory method that sets the simpleEOP parameter to true?
This parameter is now handled at a different level than before (it was called withTidalEffects or something like that). Now, it is handled directly from within the EOP interpolation code. The difference between using "simple EOP" and "complete EOP" is that during the interpolation of Earth Orientation Parameters, the tidal effects are either ignored or commputed and then added to the interpolation. This is a very long computation. Of course, it increases accuracy, but it is meaningful mainly for extremely high accuracy needs, as it is a few millimeters at Earth surface. I guess that for almost all needs, this parameter should be set to true, i.e. tidal effects should not be computed.
Beware that this can appear at several different places, as the Earth frame is for example hidden in some force models (drag, Earth potential).
Hope this helps, Luc
The 6.1 update produces similar results (differences on the order of meters, as specified in the propagator tolerance) as the 6.0 jar, but runs more slowly. Any ideas why this might be? Erin
---------------------------------------------------------------- This message was sent using IMP, the Internet Messaging Program.