Github user wesolowskim commented on the issue:
https://github.com/apache/spark/pull/14137
I added unpersists to additionally created caches and checked performance.
scc.run is slightly longer, but returned graph is cached (both vertices and
edges).
To fully optimize it and remove all unused graphs I'd have to carefully
check everything on large graph since sometimes it's easy to remove cache to
early and in this case effect will be seen after some materialization on graph
outside scc.
I hope to get it accepted in current state (that is much more performant at
least for larger graphs).
Later on I'll try to further optimize it(it is tricky to do it elegantly
because sccWorkGraph gets overwritten twice before it is materialized, and one
loop get's executed at least once).
---
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.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
