GitHub user StefanRRichter opened a pull request:
https://github.com/apache/flink/pull/2702
Improved heaptimerservice
This PR improves the performance of `HeapInternalTimerService`by pushing
all key-group related overhead into the snapshot part. This is done by
performing efficient online-partitioning w.r.t. key-group at the beginning of a
snapshot.
While this adds small computational and space overhead (both linear in the
number of registered timers), I believe this is well worth getting rid of
nested sets on timer processing.
Furthermore, this can now be extended to perform the actual state writing
asynchronously in a followup work.
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/StefanRRichter/flink improved-heaptimerservice
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/2702.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #2702
----
commit 938b4c9be8c5bd589424cb6410ae0c813546a485
Author: Stefan Richter <s.rich...@data-artisans.com>
Date: 2016-10-27T11:21:54Z
Improved performance for HeapInternalTimerService
commit 6b7389972c90dd17e801d180873dda885277a3fa
Author: Stefan Richter <s.rich...@data-artisans.com>
Date: 2016-10-27T11:33:18Z
Caching key hash codes
----
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
