<div dir="ltr">Dave-<div><br></div><div>Thanks for your reply. If it is of any help, it appears to be related to total space dedicated to gradient information. If I increase -gbs then the size that gradfil1.tmp is smaller when it fails. Depending on the size -gbs, it returns the stack identifier message and fails when gradfil1.tmp is between 2.5e6 and 3.5E6 KB. I could write the TPL more efficiently for the example dat file at the link, but the goal here is to allow more complex models via the design matrix as in glmmadmb. I'm working on including multinomial weights which will reduce the magnitude of the data and design matrices as long as numeric covariates are not used.</div>
<div><br></div><div>If there is any testing I can do on my end, please let me know.</div><div><br></div><div style>regards --jeff</div></div>