raxml ng bootstrap and pthreads

[Shoyo]

I ran a simple analysis on raxml-ng with a single gene dataset and ran the --all analysis. Opening the raxml.support tree in figtree, some bootstrap values were above 100. I ran a --search and --bootstrap analysis separately then plotted the bs values on the best ML tree using --support but still got values above 100. Am I missing something in my command?

here is my command

./raxml-ng --all --tree rand --msa 12S_alignment.phy --msa-format PHYLIP --data-type DNA --seed 1 --model TIM2{2.1821/8.5127/1.0000/15.4704}+G4{0.3420}+FU{0.4326/0.0471/0.0966/0.4237} 

and i attached the resulting .tre file

The number of threads also defaults to 4 and not 2 as specified at the beginning of the program. On my small laptop, this resulted in a termination with a seg fault 11 error but once I specified the number of threads as 2, the analysis completed.

[Alexey Kozlov]

Hi Shoyo,

thanks for reporting this!

Your command like looks fine, so it must be a bug.

By default, the number of threads is equal to the detected number of CPUs, which is often twice the number of actual
physical CPUs because of the hyper-threading. It is the same on my laptop, for instance. Although raxml-ng usually
doesn't show any speedup when using these additional "virtual" cores, it obviously shouldn't result in a segfault (and
it doesn't on my laptop, at least for all datasets I tested so far).

Could you please send your alignment file and raxml.log file to my e-mail?

Best,
Alexey

> --
> 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.

[Alexey Kozlov]

For the record: both aforementioned bugs (segfault and incorrect BS support values) have been fixed in RAxML-NG v0.5.1
available here:

https://github.com/amkozlov/raxml-ng/releases/tag/0.5.1