Nest survival file failure RFI

[birdman]

I am working on a large nest survival analysis incorporating 17 years and 2 nest stages for 34 groups. From earliest initiation to latest fledge/fail is 145 days, so I have 4896 parameters in a full "t" model (I'm not trying to run such a model, but need it if I am to assess support for T or TT models). I am also incorporating 3 locations with covariates (coded 1, 2, 3). I understand the problems inherent in using covariates as opposed to groups, but I'm afraid coding locations as groups as well would cause my machine to implode. I should mention that I've got a new machine that should have at least reasonable processing power. So, two issues for anyone with feedback.

After running a few bigger models, the ability of the program to move results into the results browser became very slow and clunky, and the file size went through the roof (i.e. 550,000 KB). Is this to be expected or should I suspect something is getting gummed up in the file? I should also note that during any such event, my computer otherwise remains fully responsive; only MARK is bogging down. Ultimately however, the program crashed and now when I try to retrieve the file, I get an error as follows:

Error Code: 11[ ]
Subsystem: BASE
Description: Dynamic Memory Low
CallStack:
DISPLAY_RESULTS:INIT (Line: 169)
MYTOPAPPWIN:BROWSERESULTS (Line: 18)
MYTOPAPPWIN:FILEOPEN (Line: 164)
MYTOPAPPWIN:OPENFILE (Line3)
APP:START (Line: 855)

Secondly, and more critically, I tried to run a TT*stage + site model using the suggested (in the Gentle Intro) eq(X,X) function to parse out two dummy variables to specify sites. Though clunky to run as described above, the model converged just fine. When I retrieve the file however, the DM is not setting up correctly, such that my eq(site,2) and eq(site,3) columns are split into two columns and I lose the right two columns of the DM, though they are correctly labeled. The Beta labels are:

Intercept stage site1 site2 Trend Trend^2 stage*TT

I modeled and should retrieve:
1 1 eq(site2) eq(site3) 1 1 1
1 1 eq(site2) eq(site3) 2 4 4
but instead get:
1 1 eq(site 2) eq(site 3) 1
1 1 eq(site 2) eq(site 3) 2

Thoughts on why it would be splitting the eq function into two columns and dropping to others? Certainly I could recode it, but I'd like to know this isn't somehow fowling (sic) my results.

Thanks, and if I can provide more/better information, just let me know. The abstract isn't due until Sunday....