search

drivenbyentropy / aptasuite

A full-featured bioinformatics software collection for the comprehensive analysis of aptamers in HT-SELEX experiments.
https://drivenbyentropy.github.io/
GNU General Public License v3.0
24 stars 11 forks source link

AptaSuite not taking advantage of multiple cores #57

Closed drthanwho closed 5 years ago

drthanwho commented 5 years ago

Hey,

I have been noticing that when Aptasuite is trying to load information (only confirmed for the clusters page so far) it is only using one of the four available cores, which results in each cluster taking 30+ mins to load. However, in a few rare occasions it has used all four cores and loaded clusters withing a couple minutes.

I was wondering if this is a bug or limitation of the program? Has anyone else experienced this?

image

drivenbyentropy commented 5 years ago

Hi,

When loading the different pages in the GUI, AptaSuite is not bound by CPU speeds but by data I/O. While I did implement this process in a threaded fashion, in my experience, it is the hard drive speed which makes most of the difference in terms of loading speed.

Do you have an SSD or a traditional HDD in you system? Given the amount of data HT-SELEX constitutes, I would recommend at least an SSD or, even better, a modern M.2 drive.

I do plan to look into further optimizing this in the future, however it might take me a while to get to it.

Hope this helps!

drthanwho commented 5 years ago

Hey,

For different pages I can see that it would be a speed problem. But looking at the different clusters I am seeing that when I have all four cores utilised loading progress is significantly faster, whereas every time it takes half an hour there is only one core utilised.

Because I would think that cores being used and data transfers are not really linked to each other.

In terms of drive though, I have the data on a HDD with a USB 3.0. Unfortunately, I don't have an SSD to test this theory. Especially, since the data AptaSuite generated are around 400GB total

Thanks for the quick response!

drivenbyentropy commented 5 years ago

Closing. LazyCache should improve the user experience related to this ticket. Please reopen if you still have an issue with this.