-
Notifications
You must be signed in to change notification settings - Fork 280
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
OverflowError in ParticleBitmap #4471
Comments
Hi @weiguangcui , thank you for reporting this First of all, I see that the line where the crash occurs was added in #4307, so it may be a regression. Can you please try downgrading yt to 4.1.4 and see if it works better ? That said, I'm surprised this crashes, given that this line is attempting to convert an |
Hi @neutrinoceros, I can confirm that yt@4.1.4 works totally fine, though it is not fully done (taken too much time to load all these particles). The error did not show up as immediately in the previous case. |
I think the It's expected that 512 overflows a 2 bytes representation. (edit: I was confusing "byte" and "bit" here, they sound identical in my mother language ! 2 bytes is however not sufficient to represent an int64 in the general case, so I think my conclusions still stand ?) I assume the correct fix would simply be to allocate more bytes to this part of the hash: 64 bits (8 bytes) would be needed to completely remove the risk of seeing that error pop up again. Other parts of the hash could also use similar patches (I see |
@weiguangcui I've opened #4472 to fix this. Can you try building yt from this branch and see if it fixes your issue ? This command should suffice to build with my patch: python -m pip install git+/~https://github.com/neutrinoceros/yt.git@fix_particle_bitmap_hash |
my patch has been accepted and will be part of yt 4.2.1, in the mean time you can install yt from source with
and of course, you are still welcome to report any problems ! thank you again ! |
Sorry for the late reply. But I can confirm that the latest version has no problem with that any more. |
Bug report
Bug summary
Code for reproduction
This is a large simulation, with 512 snapshot files and about 5*10^10 particles in the whole simulation.
Version Information
Thank you for your help.
Best,
Weiguang
edit by @neutrinoceros : added markup to the python-traceback block
The text was updated successfully, but these errors were encountered: