PointCulling as trait #273
Conversation
|
So there's good news and bad news. 👍 The good news is that everyone that needs to sign a CLA (the pull request submitter and all commit authors) have done so. Everything is all good there. 😕 The bad news is that it appears that one or more commits were authored or co-authored by someone other than the pull request submitter. We need to confirm that all authors are ok with their commits being contributed to this project. Please have them confirm that here in the pull request. Note to project maintainer: This is a terminal state, meaning the ℹ️ Googlers: Go here for more info. |
src/octree/batch_iterator.rs
Outdated
| iterator.try_for_each(|point: Point| point_stream.push_point_and_callback(point))?; | ||
| point_stream.callback() | ||
| let mut iterator: Box<Iterator<Item = NodeId>> = | ||
| get_node_id_iterator(&self.octree, &self.culling); |
There was a problem hiding this comment.
here the origin of most changes. (still wip) the node id iterator returns the id filtered by the geo query expressed by the culling and then the operation on the nodes can be parallelized.
There was a problem hiding this comment.
Where does get_node_id_iterator come from? I don't see it in the other PR.
| #[derive(Debug, Clone)] | ||
| pub struct OrientedBeam { | ||
| pub struct OrientedBeam<S> { |
There was a problem hiding this comment.
How about making this its own small PR?
There was a problem hiding this comment.
I had to change the f64 in S to allow for Sync and Send for the trait. The other changes except for the trait are again already in master, here unfortunately yet not visible cause the PR of NodeIdIterator is still open
|
At a high level, do you think we could keep the |
|
Resuming from the spontaneous pairing session with Nikolai: keeping the point culling enum struct allows to spare costly dereferencing from Arc and Box with the cost of few disambiguation lines https://play.rust-lang.org/?version=stable&mode=debug&edition=2018&gist=b5ad139bdfae3bcc3e93d15911c8ee0a |
|
this PR has been splitted up. |
To prepare parallelization across nodes, we split into node search first and then point iteration and filtering. PointCulling has being reformatted into a trait to perform the most common operations.