track cause of crash
 

Hello, I am running a case in v3.3 that keeps on crashing on me before the run is completed. Is there anyway to have the results of the last iteration before EARTH crashes saved someplace, so I can try and get a better idea of what variable is causing EARTH to crash?

Thanks in advance,

Matt

Re: track cause of crash
 

Matt,

I had a similar problem with earth. My problem was two fold.

1) The seetting under time dependence I set up incorrectly and was dumping the wrong information that cause v 3.3 to crash.

2) The other was setting up cylindrical boundary conditions.

I read through POLIS and selected the appopriate value for my model in both settings, in my case to turn off the cylindrical boundary conditions.

More detail on the type of problem you are running would help in diagnosing an exact cause of the crash.

It could be that you are running out of hard drive space, this happen to me on more than one occasion. P 3.3 consumed over 3GB of hard drive space with run time files and windows page file.

I hope this helps

Leon

Re: track cause of crash
 

If you know at what sweep it fails, you can run it again, and end the run just before that sweep by means of the interrupt facility on the graphical display. Alternatively, just set the case to run for slightly less sweep numbers than that at which failure occurs.

Another alternative: If the run is steady, set IDISPA=279 (say for example) and CSG1=P, PHOENICS will then dump a PHI file named P279 at the 279th sweep.

Any of the foregoing should get you a PHI file to look at for inspection of the field variables through PHOTON.