Author |
Message |
Mongi Mansouri (mongim_pde)
Member Username: mongim_pde
Post Number: 10 Registered: 11-2003
| Posted on Sunday, December 28, 2003 - 12:24 pm: | |
Hi all... I had posted this message earlier with the wrong sbject. Hope it does not cause confusion to anyone... I need help with the the following: the Export Tool in FlexPDE appears to be saving partial data in long-time time dependent problems... Why? I am using FlexPDE 2.22. I have just found out that the export tool in FlexPDE does not keep all the data saved to an ascii file. After a long time it begins saving data starting from a later time point and forgets about all the previous data... I don't understand this. Is this an intended feature or is it a bug? The program should keep all the data saved in text files!! After running my program for two days, I was disappointed to find that my export files did not have all the data saved. Whare did all my data go? Is there a way to tell FlexPDE to keep all the info saved in ascii files? This is really a big problem. Hope you can help... Thank you. /Mongi M.
|
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 85 Registered: 06-2003
| Posted on Monday, December 29, 2003 - 04:49 pm: | |
I don't remember the history of this very well, but at some point long ago we imposed a limit of 12228 points on a history plot. If the list grows above this value, the data are averaged back to 8192 points, and the run continues, only to be averaged again when 12228 is reached. This was intended as a fix to some user problem, and I suppose it was related to how many points one could actually display on a plot. But in retrospect it seems a really silly thing to do. I have removed this limit and posted revised versions at www.pdesolutions.com/download/xfpde222c1win.exe and www.pdesolutions.com/download/xfpde310c1win.exe. These are self-extracting archives containing only flexpde.exe (or flexpde3.exe). They are not full installations. Version 2 uses 16-bit integers in some places to store counters, so there is a possibility of trouble if the problem runs past 65535 timesteps. This should not be a problem with version 3, but neither of the posted versions has been fully tested in cases of very large numbers of time steps. In version 4, we have provided a selectable limit, with "no limit" as the default behavior. |
|