Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[bwGetOverlappingIntervalsCore] Got an error #50

Closed
zoink opened this issue Nov 4, 2017 · 2 comments
Closed

[bwGetOverlappingIntervalsCore] Got an error #50

zoink opened this issue Nov 4, 2017 · 2 comments

Comments

@zoink
Copy link

zoink commented Nov 4, 2017

Thanks for the help on the last error!

The error I get now is:
[bwGetOverlappingIntervalsCore] Got an error
An error occurred while fetching values!

I'm fetching dnase values quickly with a PyTorch dataloader (i.e. bw.values(chrm,start,end)). The error doesn't occur when I'm using the main process (num_workers=0) to fetch the values.

When I set num_workers to 8 like below then I get the error.
loader=DataLoader(valdata,batch_size=20,num_workers=8)

num_workers refers to the number of subprocesses fetching the data.
"how many subprocesses to use for data loading. 0 means that the data will be loaded in the main process (default: 0)" from PyTorch doc.

So I think the error might have to do with simultaneously fetching multiple interval values from the same open file.

Is there anyway to circumvent this? Thanks so much!

@dpryan79
Copy link
Collaborator

dpryan79 commented Nov 4, 2017

This happens frequently when using multiple threads to fetch data. The cause is usually the remote server limiting the number of concurrent connections. Consequently you're best served to minimizing the number of threads concurrently contacting the same server.

@zoink
Copy link
Author

zoink commented Nov 4, 2017

Thanks, I was able to increase num_workers once I requested a node with more cores/sockets/threads.

@zoink zoink closed this as completed Nov 4, 2017
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants