microDM / MicFunPred

A conserved approach to predict functional profiles from 16S rRNA sequence data
MIT License
6 stars 6 forks source link

Possible Memory Leak #8

Open ch16S opened 2 years ago

ch16S commented 2 years ago

Hey guys,

I have successfully run MicroFunPred with the test data set, but when I try to run a large set (>1000 samples >3000 taxa) my memory usage creeps up slowing until the process gets killed due to maxing out the memory usage (after 24 hours or so). I was running 16 GB of Ram with 8 CPUs on UBUNTU 20.01.

Is this expected behavior or should memory usage be lower?

Cheers, Chris

dattamongad commented 2 years ago

@ch16S It is quite unusual. Because we had tested MicFunPred on large datasets and the only process which requires RAM is loading reference database (~6 to 8 GB). Can you share your data with me and the command you have used at datta@microdm.net.in?