nolanlab/citrus

More a Question than an Issue: Is there a multicore option for citrus.predict.classification

Opened this issue · 1 comments

Hi,
I could not find the answer within the documentation to the following question. If I overlooked something, please be so kind to point me there.
You provided comprehensible full example for citrus.predict.classification. With my data it takes a long time to compute. I was wondering, is there a multicore approach for this example?
Especially the clustering step takes a very long time to finish.

Best Regards
Michael

Hi Michael,
The best way to accelerate runtime is to compile Rclusterpp with multi-threaded support as it is the most computationally intense step (as you have noticed!). There are some old instructions on how to do that here: https://github.com/nolanlab/citrus/wiki/Installing-Citrus#Citrus_Installation_on_Mac_OS_X_using_MacPorts_and_g48_With_OpenMP_Support
Unfortunately, they might be a bit out of date, but it's worth giving it a shot!