-
Notifications
You must be signed in to change notification settings - Fork 311
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
[BUG] PageRank Values Miscalculated with Float Weights #1330
Comments
Thanks for reporting this. Sounds like a bug as the current single GPU version should ignore weights. We are looking into it. One option is to use the new multi-GPU version which uses edge weights if the input graph is weighted. #1313 (comment). |
Also, we are going to connect the upgraded PageRank backend (which accepts weights) to the single-GPU API in this release (0.18). This bug should go away as we do so. |
Glad to hear it! Thank you for looking into this so quickly. |
Describe the bug
When using
cugraph.pagerank
on a graph initialized withfloat
values for theedge_attr
parameter, the returned values are well outside the [0, 1] range expected by the PageRank calculation (often by many orders of magnitude). If the underlying values are instead ofint
type, the calculation works as expected. The function is currently supposed to ignore the weights column for the calculation, so it seems to purely be a bug.Steps/Code to reproduce bug
Output:
Expected behavior
The returned PageRank values should all be in [0, 1], and presumably match when the weights are
1
and1.0
.Output:
Environment overview (please complete the following information)
Environment details
Please run and paste the output of the
cugraph/print_env.sh
script here, to gather any other relevant environment detailsClick here to see environment details
The text was updated successfully, but these errors were encountered: