Github user StephanEwen commented on the pull request: https://github.com/apache/flink/pull/1093#issuecomment-138484103 Would be interesting to benchmark the off-heap mode for memory intensive jobs. I would tend to agree with your performance assessment. Since Flink always did do off-heap style techniques (albeit on heap) the difference should not be gigantic. I'd be happy to if you took this over. I could merge it (it breaks nothing) and you could pick up the mini cluster and script configs.
--- 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. ---