Hi to you both, and thanks for this great package. I'm finding it very useful.
I'm in the process of incorporating phylolm in the phylopath package for a couple a reasons, one of which is the performance on large trees. A slow performing part of the function is, however, if one wants to estimate confidence intervals using the boot argument (which I presume is the recommended method for both phylolm and phyloglm). This seems like an "embarrassingly parallel problem" to me and I was wondering whether you have considered to include parallel support for this?
It would basically involve parallelizing the for (i in 1:boot) { } part of the two functions, and should be relatively straight-forward. I would be happy to prepare a pull request for that, but it may take me a couple of weeks to get to it, since I'm nearing the final stages of my thesis work.
Hi to you both, and thanks for this great package. I'm finding it very useful.
I'm in the process of incorporating
phylolm
in thephylopath
package for a couple a reasons, one of which is the performance on large trees. A slow performing part of the function is, however, if one wants to estimate confidence intervals using theboot
argument (which I presume is the recommended method for bothphylolm
andphyloglm
). This seems like an "embarrassingly parallel problem" to me and I was wondering whether you have considered to include parallel support for this?It would basically involve parallelizing the
for (i in 1:boot) { }
part of the two functions, and should be relatively straight-forward. I would be happy to prepare a pull request for that, but it may take me a couple of weeks to get to it, since I'm nearing the final stages of my thesis work.Thanks again.