Error with swapping chains?

[Ziv Lieberman]

Hi all,

I've been having some issues getting ExaBayes to run recently in a few ways. I'm using the CIPRES gateway to access ExaBayes 1.5.1.

The first issue arose when the MCMC sampled the first generation, then seemingly produced no outputs whatsoever over the next 24 hours. I saw that there was a warning (" WARNING! The number of processes and number chains run in parallel are not a multiple of the overall number of processes. Parts of the code may be executed less efficiently."), so I adjusted the number of couple chains to be appropriate.

On running the analysis again, the MCMC proceeded to generation 500, then failed. The stderr has the following messages:

exabayes: ./src/mcmc/CoupledChains.cpp:383: void CoupledChains::executePart(uint64_t, uint64_t, ParallelSetup&): Assertion `_chains[remoteId].getGeneration() == swap.getGen()' failed.
exabayes: ./src/mcmc/CoupledChains.cpp:383: void CoupledChains::executePart(uint64_t, uint64_t, ParallelSetup&): Assertion `_chains[remoteId].getGeneration() == swap.getGen()' failed.

Please see the full stderr.txt attached. I am looking for any insight on how to ameliorate this issue. 

I have previously run ExaBayes successfully using more parallel processes, more heated chains, on a larger data set with more partitioning, so I intuitively feel that the data amount or complexity is not a major contributor here.  It may be notable that my prior successful analyses were on version 1.5.0.

I can provide alignment and partition stats or even the whole files if necessary. For now I have attached the stderr, stdout, and the config file used.

Thank you!

-Ziv

[Alexandros Stamatakis]

Dear Ziv,

The student who developed ExaBayes unfortunately left for industry about
10 years ago already.

Please send me the input files and I will see what I can do.

Alexis

> --
> You received this message because you are subscribed to the Google
> Groups "ExaBayes" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to exabayes+u...@googlegroups.com
> <mailto:exabayes+u...@googlegroups.com>.
> To view this discussion visit https://groups.google.com/d/msgid/
> exabayes/1f5ea8bd-85de-497e-8e30-547673e326cbn%40googlegroups.com
> <https://groups.google.com/d/msgid/
> exabayes/1f5ea8bd-85de-497e-8e30-547673e326cbn%40googlegroups.com?
> utm_medium=email&utm_source=footer>.

--
Alexandros (Alexis) Stamatakis

ERA Chair, Institute of Computer Science, Foundation for Research and
Technology - Hellas
Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology

www.biocomp.gr (Crete lab)
www.exelixis-lab.org (Heidelberg lab)