search

cliffren / PENCIL

PENCIL is a novel tool for singlecell data analysis to identify phenotype associated subpopulations and informative genes simultaneously.
GNU General Public License v3.0
25 stars 1 forks source link

How to speed up the process using CPU？ #9

Open Payne-Wang opened 11 months ago

Payne-Wang commented 11 months ago

Hello, thank you for providing such a great analysis tool. I have a question to ask. Our server here does not have a GPU installed, and we are currently running on CPU only. However, I've noticed that the speed is quite slow. I used the dataset from tutorial 1 and ran the code according to the tutorial, but it took more than 12 hours. At the same time, I ran another dataset with 10,000 cells, and it has been running for over 12 hours as well. Is there a method to speed up the process using CPU? I would greatly appreciate your help.

cliffren commented 9 months ago

Thank you! But unfortunately, this model is entirely based on the pytorch framework, and trying to achieve acceleration on a cpu is very difficult indeed. Having a GPU makes things much better. Sorry, there doesn't seem to be much we can do at the moment.