MARK crashes when opening results file

[Morten Frederiksen]

This one is new for me: when I try to open a certain results file, MARK immediately crashes, with the usual impenetrable error message. I last accessed the file a few hours ago, and closed MARK down in orderly fashion. There is no problem opening other results files ... I have uninstalled and reinstalled MARK, to no effect. Does anybody have a clue to what's going and what I could do to resolve the problem? Not being able to access results of a major analysis is a pain ...

Morten

[rlong]

I've had similar experiences when dealing with fairly large model sets. I found that if I created a new subdirectory on my machine, copied the MARK files from their original directory into the new one (in my case this included the .dbf, .fpt, and .inp files), MARK would then open the results file from this new location without incident. Good luck!

[gwhite]

Morton:
There is a bug in the versions of MARK compiled with the Visual Objects 2.7 compiler that may be your problem. "Large" output files (many large PIMs, long design matrix, lots of parameters) that generate a large VC matrix will cause a memory error when you try to re-open the results file. I've been working with the Visual Objects folks to fix this problem, but no progress because I can't duplicate it with a really clean example. However, the older versions of MARK will open the files just fine.

The comment by Robert Long that copying the files to a new subdirectory is a new one on me. That hasn't worked in the past with the problem I describe above. The only way I could do anything with the files was to use an older version of MARK, which I can provide if necessary.

Gary

[Morten Frederiksen]

Gary,

I think my problem is exactly what you describe; the model I had run most recently (Burnham live/dead) had 43 occasions and 1105 parameters ... If you could supply an older version of MARK, that would be very helpful. I probably have a really old version somewhere (~2000), but I don't suppose I have to go that far back.

Morten

[tinah]

I had the same problem where MARK all of a sudden crashed because I was running large models with long design matrices, but when I installed MARK 4.0, I still got an error telling me there was no MARK results database, which is clearly in the folder I created for the analysis. Do I need to restart this analysis or is there something else I can try? I really do not want to have to rebuilt and rerun the ten models I have already run.