Github user jkbradley commented on the pull request:
https://github.com/apache/spark/pull/2868#issuecomment-61153266
I got similar results on 16 nodes using MNIST8m; basically no change in
runtime (+/- a few percent at most). But those tests were for shallow trees.
I worry that this patch will only help for depths at which we might as well
switch to local training. In tests with Sequoia Forests, do you know how much
distributed node ID caching helps? I wonder if node ID caching will only be
helpful for local training.
Running more tests...
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
