Author |
Message |
Mongi Mansouri (mongim_pde)
New member Username: mongim_pde
Post Number: 1 Registered: 11-2003
| Posted on Wednesday, November 26, 2003 - 03:07 pm: | |
In a time dependent heat problem, I wanted to run the program for a long time (till it reaches a steady state). The program behaves well until the pgx file becomes about 100 MB and then it stops running and saving results. I am only sending two xy-plot widows to the pgx output file. I am not sure why this happens... Should I reinstall? Please help. Thanks. /Mongi |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 67 Registered: 06-2003
| Posted on Wednesday, November 26, 2003 - 05:20 pm: | |
I don't have any record of a bug like this in version 2, but it's been a long time since I used that version. You could try downloading and installing www.pdesolutions.com/download/versions/xfpde222b1win.exe This is a self-extracting archive containing only the executable flexpde.exe, the latest modification that was made to version 2 (1/15/2003). Change the name of your existing file (so you can back up if you have to) and extract the new one in its place.
|
Mongi Mansouri (mongim_pde)
New member Username: mongim_pde
Post Number: 2 Registered: 11-2003
| Posted on Sunday, November 30, 2003 - 06:40 pm: | |
Dear Robert... I have uninstalled and reinstalled FlexPDE. I ran my program again and the same thing happened. ie, the program will run up to a cartain point and then halts. It appears to be stuck and practically hugs all the CPU time. I had to force-stop it using "Task Manager" in Windows. There is no error message issued and this makes it hard to see what the problem is. I am working on a simplified version of the code to send it to you. Hope you will be able to help. |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 70 Registered: 06-2003
| Posted on Monday, December 01, 2003 - 10:22 pm: | |
I dimly recall having a problem like this early in the development of version 3, but there is no record of the fix. I believe it had to do with the fact that FlexPDE tries to supress monitors if the display time is large compared to the compute time. This supression depends on the microsecond clock, and can get into difficulty with roll-over at long times. If this is the problem, it should affect only the monitors, and not the execution or plot files. On the assumption that this was the cause, I copied the version 3 time clock routine into version 2, and put the new code at www.pdesolutions.com/download/version/xfpde222b2win.exe. Give this one a try, and see if it fixes the problem. |
Mongi Mansouri (mongim_pde)
Junior Member Username: mongim_pde
Post Number: 3 Registered: 11-2003
| Posted on Tuesday, December 02, 2003 - 09:57 am: | |
Dear Robert... I ran my pde code again with the new fix for FlexPDE 2.22 that it did not change anything. The program simply stops going at exactly the same point in time, issues no errors and goes into a "Not Responding" process. I understtod your explaination for this problem. Could be due to "Operating System"? I am using W2KPro on a P4 PC. Should I change anything about the display settings? What do you suggest? Thanks very much... |
Robert G. Nelson (rgnelson)
Moderator Username: rgnelson
Post Number: 71 Registered: 06-2003
| Posted on Tuesday, December 02, 2003 - 05:54 pm: | |
I have a new theory. Version 2 uses a 16-bit number to store the count of plot times. I guess I never expected more that 65536 plots. In any case, it appears that the crash occurs on the 65537th plot time (maybe - I can't tell exactly). I assume you are using the plot time as a limit to the timestep, but you may have to increase the plot interval to get a longer run time. I will look into changing this, but maintaining version 2 is becoming difficult because of evolving libraries, etc. (Theoretically,it was frozen two years ago.) |
Mongi Mansouri (mongim_pde)
Member Username: mongim_pde
Post Number: 4 Registered: 11-2003
| Posted on Wednesday, December 03, 2003 - 10:15 am: | |
Dear Robert... Thank you very much for your help. The latest patch you provided regarding the plot count in FPDE 2.22 has fixed the problem. Your diligence and proper thinking about FlexPDE users concerns are greatly appreciated... /Mongi |