Skip to content
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

[k-NN] Avoid additional copy to stream during binary doc values deserialization #1736

Closed
Tracked by #1709
jmazanec15 opened this issue Jun 10, 2024 · 2 comments
Closed
Tracked by #1709
Assignees
Labels
Enhancements Increases software capabilities beyond original client specifications indexing-improvements This label should be attached to all the github issues which will help improving the indexing time. k-NN v2.16.0

Comments

@jmazanec15
Copy link
Member

jmazanec15 commented Jun 10, 2024

Description

Related #1709

During deserialization, we have an intermediate step where we copy a bytesref to a bytesstream. This can be removed as it adds some overhead and does add any benefit. With #1087, this will be no longer necessary. But until then, it could add some benefit.

From the experiments in #1709 , the diff with and without the optimization is:

Results - w/o change

Run # p50 latency (ms) p90 latency (ms) p99 latency (ms) Recall
1 674 684 692 0.99998
2 674 684 692 0.99998

Results - w/ change

Run # p50 latency (ms) p90 latency (ms) p99 latency (ms) Recall
1 568 584 594 0.99998
2 568 584 596 0.99998

PoC commit: jmazanec15@7984618.

@jmazanec15 jmazanec15 added Enhancements Increases software capabilities beyond original client specifications and removed untriaged labels Jun 10, 2024
@navneet1v navneet1v added the indexing-improvements This label should be attached to all the github issues which will help improving the indexing time. label Jun 11, 2024
@navneet1v
Copy link
Collaborator

This intermediate step of copy byteref to bytestream is also present when we read vectors during merge. Hence marking this issue as indexing-improvements too.

@vamshin vamshin added the v2.16.0 label Jul 2, 2024
@vamshin vamshin changed the title Avoid additional copy to stream during binary doc values deserialization [k-NN] Avoid additional copy to stream during binary doc values deserialization Jul 15, 2024
@vamshin vamshin added the k-NN label Jul 15, 2024
@navneet1v
Copy link
Collaborator

@jmazanec15 can we close this GH issue as the feature is added in k-NN plugin,

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Enhancements Increases software capabilities beyond original client specifications indexing-improvements This label should be attached to all the github issues which will help improving the indexing time. k-NN v2.16.0
Projects
Status: 2.16 (First RC 07/23, Release 08/06)
Status: Done
Development

No branches or pull requests

3 participants