Raxml does more starting tree reps than requested.

[Chris]

Hi,
I am using version 8.1.17.
I started raxml with MPI:

mpirun -np 22 -wdir /absolute-path-to/tree-search/ raxmlMPI_8.1.17 -m GTRGAMMA -p 123 -q partitions.txt -s supermatrix_Gb.phy  -n Tmr -N 30

It already computed MP starting trees until the ending ".41" and has finished by now 36 reps.
Did I miss something or why is it computing more reps than requested?

Best
Christoph

[Alexandros Stamatakis]

Hi Christoph,

It does indeed compute a bit more trees than requested, this is
intentional, if I remember well it should compute a total of 44 trees,
that is a multiple of 22 larger than the number of trees you specified.

Alexis

> --
> You received this message because you are subscribed to the Google
> Groups "raxml" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to raxml+un...@googlegroups.com
> <mailto:raxml+un...@googlegroups.com>.
> For more options, visit https://groups.google.com/d/optout.

--
Alexandros (Alexis) Stamatakis

Research Group Leader, Heidelberg Institute for Theoretical Studies
Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
Adjunct Professor, Dept. of Ecology and Evolutionary Biology, University
of Arizona at Tucson

www.exelixis-lab.org

[Chris]

Hi Alexis,

thanks. It did indeed 44 replicates. I assume they are all considered for determining
the best tree.

Best
Christoph

[Alexandros Stamatakis]

On 22.04.2015 23:58, Chris wrote:
> Hi Alexis,
>
> thanks. It did indeed 44 replicates. I assume they are all considered
> for determining
> the best tree.

yes they should :-)

alexis

> > an email to raxml+un...@googlegroups.com <javascript:>
> > <mailto:raxml+un...@googlegroups.com <javascript:>>.

> > For more options, visit https://groups.google.com/d/optout

> <https://groups.google.com/d/optout>.

>
> --
> Alexandros (Alexis) Stamatakis
>
> Research Group Leader, Heidelberg Institute for Theoretical Studies
> Full Professor, Dept. of Informatics, Karlsruhe Institute of Technology
> Adjunct Professor, Dept. of Ecology and Evolutionary Biology,
> University
> of Arizona at Tucson
>

> www.exelixis-lab.org <http://www.exelixis-lab.org>

[taua...@gmail.com]

Hi Alexis,

Sorry to reopen this. I also noticed my runs are calculating more than the 10 trees I asked for. Why is that? And how can I know how many it will calculate (so maybe I ask for less for a total of ten)?

It did not happen in the Pthreads version, but it is happening now that I am using the Hybrid version 8.2.10

srun SLURM_OPTIONS raxmlHPC-HYBRID-SSE3 -T 4 -N 10 -p 1313$SLURM_ARRAY_TASK_ID -s $file -m PROTGAMMALG4X -n ${outfile}

Thank you!

[Alexandros Stamatakis]

hi there,

it only happens with MPI (there should be some previous posts about
this) it does slightly more trees than requested, basically it rounds up
the number of inferences so that there is an identical number of
inferences per MPI process, so if you want to infer 10 trees with 4 MPI
processes it will actually do 12 as 4 * 3 is 12 :-)

Alexis

www.exelixis-lab.org

[taua...@gmail.com]

Awesome, thanks, Alexis. I changed to 5 MPI processes and now it is calculating exactly 10.

Tauana