FEAP User Forum
FEAP => Programming => Topic started by: Enginit on July 09, 2021, 11:11:20 AM
-
Dear all,
I create Istart file with 10000 simulations (include Icir-ss-1 ... include Icir-ss-10000 ... STOP).
FEAP runs around 173 simulations and then stops with this error:
At line 65 of file pinpfl.f (unit = 12, file = 'Scir-ss-38038.edis0000')
Any suggestion would be a great help.
Many thanks
P.s: I am using Feap v8.4 and working on user-defined element.
-
Is there anything more to the error message? Like an actual identified error issue?
Further what is on line 65 in pinpfl.f? In my version of 8.4 it is a blank line.
-
Does this problem run by itself?
I have a vague recollection that there was a bug somewhere -- but it was a long time ago so I do not remember details.
-
A 0000 file should not occur -- should be something between 0001 and 9999
-
Is there anything more to the error message? Like an actual identified error issue?
Further what is on line 65 in pinpfl.f? In my version of 8.4 it is a blank line.
Dear Prof. Govindjee,
I would like to thank you for your help.
1. The error message is as followings:
At line 65 of file pinpfl.f (unit = 12, file = 'Scir-ss-29029.edis0000')
Fortran runtime error: End of file
2. In think my Feap version is 8.4a. The line 65 in pinpfl.f is in the attached figure.
Best regards,
-
Does this problem run by itself?
I have a vague recollection that there was a bug somewhere -- but it was a long time ago so I do not remember details.
Dear Prof. Taylor,
I would like to thank you for your help.
When I finish setting up Istart file, I run feap with Istart file as usual. It solves some files and then stops with the above errors. The output file (displacement and reaction) shows that feap makes a new file name by duplicating the number (at the end of file name) in some manner. I think there could be a problem with the subroutine which writes out the file name ... but I am so sure.
Best regards,
-
What is in file Scir-ss-29029.edis0000?
-
What files exist in the directory after the crash?
How many dof's does your user element have?
Are all the files the same except for parameters? If so, post one of the files (if not too big).
-
What is in file Scir-ss-29029.edis0000?
Dear Professor,
I also don't know what it is since I only export displacement and reaction.
Thank you and best regards,
-
This file is created by feap from a record in the input file "edis" for edge displacements.
Look to see if you have result files Ocir-ss??? Note how many trailing numbers exist if 3 digits, 001, 002, ...., you can only to 999 problems before it crashes anyway.
-
What files exist in the directory after the crash?
How many dof's does your user element have?
Are all the files the same except for parameters? If so, post one of the files (if not too big).
Dear Professor,
1. Files in the directory: elmnt.f, uallocate.f, pelmin.f(to read additional data to user element), makefile, Istart, all Ifiles, Ostart, Lstart, S file (I don't know why it shows out), .dis file, .sum file (see attachment), Ocir files
2. User element has 3 degree of freedoms (a beam element)
3. Yes, that's right. I just change the Young's modulus and Poisson's ratio for each input file, the rest of the input file is totally the same. (see attachment)
Thank you and best regards,
-
Please post Istart
-
Please post Istart
Here it is, Professor.
Thank you and best regards,
-
(1) I am surprised that your version of the code is not printing an error to the screen "Exceeded limit of multiple files (PCONTR)"
(2) It should have run to completion (even if it printed Exceeded... message), though the file names will have gotten a bit scrambled from the way they should have been. Perhaps your version 8.4a has a bug in it? It is quite old and I do not have a way to check that particular version; I checked in v8.6 If you upgrade to the current version, it will run your problem.
(3) If you do upgrade, it will run your problem. But FEAP does currently have a limitation of 999 included problems, something we can consider increasing in the upcoming v8.7 release. Note, the reason 8.6 will run your problem has to do with the fact that each of your included files has a unique name (something we do not require when running mulitple problems out of one file).
Update: I have checked that v8.5 and v8.4 will run your problem with 10000 include files. The v8.4 version I tried is v8.4.1d (01 January 2014). The code prints the Exceeded... message but continues on to completion. You can try downloading a clean copy and trying again.
-
(1) I am surprised that your version of the code is not printing an error to the screen "Exceeded limit of multiple files (PCONTR)"
(2) It should have run to completion (even if it printed Exceeded... message), though the file names will have gotten a bit scrambled from the way they should have been. Perhaps your version 8.4a has a bug in it? It is quite old and I do not have a way to check that particular version; I checked in v8.6 If you upgrade to the current version, it will run your problem.
(3) If you do upgrade, it will run your problem. But FEAP does currently have a limitation of 999 included problems, something we can consider increasing in the upcoming v8.7 release. Note, the reason 8.6 will run your problem has to do with the fact that each of your included files has a unique name (something we do not require when running mulitple problems out of one file).
Update: I have checked that v8.5 and v8.4 will run your problem with 10000 include files. The v8.4 version I tried is v8.4.1d (01 January 2014). The code prints the Exceeded... message but continues on to completion. You can try downloading a clean copy and trying again.
Dear Prof. Govindjee and Prof. Taylor,
I tried again by changing the input file name as I1, I2 ... I10000 then now feap runs all simulation successfully.
The only problem rests in the naming style of output files (see attachment).
If it is possible, I would like to complete this topic.
I would like to thank you very much for your kind suggestion and help.
Best regards,
-
Great. If the naming scheme you are seeing annoys you, you can change it in program/pnewprob.f. Just comment out the lines around line 301 that set values in positions i:i+2 in the fplt string. As long as your include,Ixxxx files all have unique names, you will safely retain all your computations in uniquely named output files.
-
Dear Prof. Govindjee,
I would like to thank you for your helpful suggestion.
Best regards,