Batch neighbour retrieval in single server case #21862
Open
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
6b721a2 to
b3150f6
Compare
…trieval-single-server-case
mchacki
reviewed
Jul 22, 2025
There was a problem hiding this comment.
Good Code in general, I think this is a step in the right direction.
But as you noted it will not yet solve the problem, as it still iterates over all neighbors, it just adds context switches for now. (You want to use those to return so that is good 👍)
I am starting to think about this on a more high level.
Is the intermediate copy into the std::vector actually necessary?
Or should we aim for an iterator approach directly. The iterator we could just keep open.
And only consume the batch we want.
If we cache then the iterator can be feed from the cache on next round.
arangod/Graph/Providers/SingleServer/SingleServerNeighbourProvider.cpp
Outdated
Show resolved
Hide resolved
- add a use-cache plug to SingleServerNeighbourProvider because for some usages - e.g. a simple tree traversal - the cache will not be useful and just waste memory. The plug is not yet used. - reverse batch size memory batch for batch vector for more efficient memory usage - get batch size from ExecutionBlock::DefaultBatchSize instead of using a "magic number"
…our-retrieval-single-server-case
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This is the first PR to retrieve neighbours in batches in traversals. This can drastically improve limited traversal runtimes on graphs with supernodes.
In the graph's
SingleServerProvider, this PR adapts how neighbours of a specific vertex are read in theexpandfunction. It now introduces a neighbour provider that is responsible for providing the neighbours in batches (which extracts a lot of code in theSingleServerProvider). For now we loop over all batches to get all neighbours inexpand. In a later PR (when the cluster case is also implemented),expandshould return one batch per call.The neighbour provider is set to a specific vertex and provides one batch of neighbours per call to its
nextfunction. Internally, it saves read neighbour batches to a cache. If the neighbour provider is set to a vertex for which the cache includes already all neighbours, the neighbour provider provides these cached batches instead of reading them again from memory.