-
-
Notifications
You must be signed in to change notification settings - Fork 404
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
Fast initialization for edge connectivity #34123
Comments
New commits:
|
Commit: |
Dependencies: #33839 |
comment:3
Can you add more comments to better explain the steps. You define |
Changed keywords from none to GSoC2022 |
comment:4
Yeah, I can add more comments. Right, I should be using a set, totally missed that. By the way, I had been trying to make this ticket's diffs show just my commits, excluding yours from #33839, but I haven't figured out how yet. |
Branch pushed to git repo; I updated commit sha1. This was a forced push. New commits:
|
comment:7
I'm not sure where I'm going wrong. My DFS initialization is not fast. In fact, it's slower than the implementation without it. Here are the results of some tests carried out on complete digraphs: Without initialization 300 vertices 500 vertices 1000 vertices Mine, with initialization 300 vertices 500 vertices 1000 vertices |
comment:8
I'm obviously missing something. Not sure what :/ |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:10
Could be because your method is recursive. Can you try a non-recursive DFS ? Also, have you checked that your method is valid with other types of graphs ? |
comment:11
Okay. I'll try to implement an iterative version.
Just did. It works on the Peterson graph, fails on Regular graphs. I get the error message " ValueError: did not find the right edge". I'll try to fix this and do more extensive testing |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:13
My latest commit removed the G_minus_T function, and instead updates the current k-1 intersection inline. I expected this to speed up the code a little bit, but that doesn't happen. Instead, it's slower than before. I'm confused Test results, which are slower than that reported earlier: 500 vertices CPU times: user 5.83 s, sys: 4.03 ms, total: 5.84 s Wall time: 5.84 s 499 1000 vertices CPU times: user 53.4 s, sys: 158 ms, total: 53.5 s Wall time: 54.1 s 999 |
comment:14
Also, I haven't yet figured out how to get the code to work on regular graphs. The " ValueError?: did not find the right edge" error stems from the fundamental cycle step, and is linked to the my_depths, my_parent and my_parent_edge vectors. I attempt to fix this by updating these vectors while doing the initalization DFS, compute_dfs_tree(). But this leads to an infinite loop in some other DFS call, update_parents_dfs(). |
comment:15
You never set any edge in |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:17
Somehow, the code is even slower: '' 1000 vertices I'll get to the bottom of this! |
Branch pushed to git repo; I updated commit sha1. New commits:
|
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:50
I did some minor review edits. Do you agree with these changes ? |
comment:51
Looks good to me! |
comment:52
So then. |
comment:53
See patchbot:
|
comment:54
Hmm. This is weird. I've checked, and can't seem to find where the error is stemming from. I even tried looking up other tickets where this error showed up (e.g #33971 comment:21) to see how they fixed it, but that doesn't help me here :/ |
comment:55
can you revert this change - EXAMPLES:
+ EXAMPLES:: |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:57
This should fix the issue. Let's wait to know if the patchbot is happy or not. |
Branch pushed to git repo; I updated commit sha1. New commits:
|
comment:59
The docbuild issue is now fixed but the patchbot has reported an error on the doctests using |
comment:60
The patchbot is now happy. |
Changed branch from public/graphs/34123_dfs_init to |
Related to #33839
Depends on #33839
Component: graph theory
Keywords: GSoC2022
Author: Enjeck Cleopatra
Branch/Commit:
8340185
Reviewer: David Coudert
Issue created by migration from https://trac.sagemath.org/ticket/34123
The text was updated successfully, but these errors were encountered: