[R-sig-ME] Resume terminated lmer fit if verbose=TRUE?

Ben Bolker bbolker at gmail.com
Wed Jul 27 15:29:12 CEST 2011


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 07/27/2011 09:16 AM, Mike Lawrence wrote:
> Hi folks,
> 
> I ran a binomial lmer on a large data set with lots of levels of a
> random effect (1000+) and a large fixed effects structure (7 variables
> all interacting). My local machine wasn't up to the task (it needs
> about 16GB of memory so far as I can tell), so I put it on a serial
> node on a supercomputer to which I have access. I asked for 100 hours
> of compute time, but it seems that the model went over time and the
> process was terminated by the system's queue. However, I ran it in
> verbose mode and have all the output, so I'm wondering if there's any
> way to use this information to resubmit the job and have it resume
> where it left off. Any ideas?
> 
> Cheers,
> 
> Mike
> 

  I'm afraid you're probably stuck.
  At the moment lme4 doesn't (I think) take starting values (this would
be a good idea in general, one could try to run with something cheaper
and get close to the answer before starting glmer), it would probably
take at least 100 hours of *elapsed* time (i.e. time during which the
computer could be working on a new version of your job) to get that code
written ...
  And yes, checkpointing would be a good idea too.

  Ben Bolker

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.10 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iEYEARECAAYFAk4wEqgACgkQc5UpGjwzenNg+QCeNULj+pqeDSN6kIjzSGtt0QKV
M+cAn3Qab0Yxeo36TvqHyzvyLLq+k1Dp
=KK5B
-----END PGP SIGNATURE-----




More information about the R-sig-mixed-models mailing list