ctlab / fgsea

Fast Gene Set Enrichment Analysis
Other
379 stars 67 forks source link

Parallel backend running out of resources #16

Closed topherconley closed 6 years ago

topherconley commented 6 years ago

Hi,

First, very nice package.

If I run fgsea hundreds or thousands of times by means of an lapply or foreach::foreach, either on linux or Mac, it generates a lot of zombie processes and then runs out of resources. It would be nice to turn off the BPPARAM or nproc option and run serially. Even better, I can then wrap this inside a solid and stable backend like doParallel combined with an iterator like foreach and not have to deal with an unsatisfactory BiocParallel wrapper.

Why I am running this hundreds or thousands of times has to do with a specific application not (historically) common in bioinformatics data sets.

Thanks in advance, Chris

assaron commented 6 years ago

Hi Chris,

I've asked this question on Bioconductor support site: https://support.bioconductor.org/p/104702/

In the meantime, you can set nproc option to 1. This shouldn't create additional processes.

Best, Alexey

assaron commented 6 years ago

As Martin Morgan said on the Bioconductor Support, this issue is fixed in R-devel version. You can try to install it from source or wait for the release.