INTERNAL PROGRAMMING ERROR

[narasimhamurthy 0]

Hi,

 

i am performing buckling analysis but getting error as show in below.

 

*** INTERNAL PROGRAMMING ERROR ***

   in file "xdsemx_i8.F", at location # 262.

 

       ioerr(1) =          1

 ****  ABORTING RUN DUE TO AN INTERNAL ERROR  ****

 

i am using Optisrtuct 2018 version

 

how to solve this error?

 

 

Thanks and Regards,

Narasimhamurthy

 

M-NG-FR-end.out

[Rahul R 316]

Could you just add PARAM,LGDISP,1 and run again.It seems this was fixed in the version(2018.0.1) which you are currently using.

[narasimhamurthy 0]

Defined PARAM, LGDISP, 1

but getting same error.

Please find attached output file.

M-NG-FR-end.out

[Rahul R 316]

Could you also add PARAM, KGRGD, YES as per the warning mentioned in the .out file. If it gives an error, then please share .out and .stat file.

[narasimhamurthy 0]

getting same error again.

M-NG-FR-end.out M-NG-FR-end.stat

[Rahul R 316]

Please remove special character (- _ . % etc)used in part name/file name/user name & also change the run folder.Run again and let us know the outcome.

[narasimhamurthy 0]

not yet solved.

 

Please find attached output file and stat file.

 

MNGFRend.out MNGFRend.stat

[Rahul R 316]

Can you give a final try by unchecking output request or param cards defined.In the past we have seen this programming error is mainly observed because of output request /param.

[narasimhamurthy 0]

removed all PARAM and output request.

 

but error not yet solved.

 

MNGFRend.out MNGFRend.stat

[narasimhamurthy 0]

is there any solution to solve this error?

[Rahul R 316]

On ‎5‎/‎31‎/‎2019 at 6:57 PM, narasimhamurthy said:

removed all PARAM and output request.

 

but error not yet solved.

 

MNGFRend.out 11.84 kB · 2 downloads MNGFRend.stat 8.81 kB · 0 downloads

Last shared files are based on 2017.2.So I am commenting based on the files shared in this quote using 2018.0.1

 

This is related to some IO issue during a matrix vector multiplication in Lanczos. I am afraid I will need the model to see what could be the problem but as a workaround you could try using MUMPS as the lanczos solver via the solvtype card and see if it works out. Alternatively you could try ddm mode (even with 1 proc).