Hi Spirit,
your output file is really huge!
I suggest you 2 possible solutions :
1. run the solution again with less output request in
   subcase select. E.g. you may sweep the request of
   SPCFORCE if you really do not need it.
2. If you use PATRAN with version 9.0 or later, then you
   may use XDB format instead of .op2 for data output. To
   my know, the XDB format is faster then op2, because it
   doesn't load all the file/data in to the memory.

Hope it would help you

regards

Hello Spirit,
I would recommend the two suggestions by zuardy, especially "attaching" the xdb file.  This will not read the results into Patran, merely reference to it.  Also, I would like to suggest two other options.
  
3. If you have more than 1 subcase in the run, lessen the number of subcases
4. Request results for only subsections of the model you may be interested in.  Make multiple runs requesting output for different subsections to verify fidelity of the entire model.

The results file sounds awfully large!

-George

OK pals, all that you were saying sounds very correct and I guess handling 4GB output data is not a clever way of working... my concern is about reading the results WITHOUT having to run again the calculation, since it takes a lot of time that I haven't. So, anybody grabs the challenge ? Thanks anyway, I fear mine is a hopeless situation...

'Ability is 10% inspiration and 90% perspiration.'