Github user shijinkui commented on the pull request:
https://github.com/apache/spark/pull/3866#issuecomment-68969847
@ankurdave thanks for reviewing :)
1. Activeness is from `innerJoin`'s result by `vprog` the message of pre
iteration.`isAttrChanged` is from solution1. It can run, but the key problem
is after 100 or more iteration, it's progressive increase one second every
iteration. So, I consider the solution 2. Solution 2 is surprised on both
performance and code logic is more clear.
2. take k-core algorithm for example. The coreness from 4 to 3 need
iterating 68 times, from 3 to 2 need hundreds. But the sum coreness of 4, 2,
1 in the final result is not important, or user can ignore the effect of
lacking computing coreness `4, 3, 2,1`. So they want't to finish the loop ahead
of time.
`isTerminal ` provide a custom condition of terminal by current iterate id
`curIter` and `messageCount`.
Default `isTerminal` is true.
3. making `initialMsg` an `Option`, as sometimes the initalMsg is empty.
when `initalMsg` is empty, aggregate message on every Vertex directly.
---
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]
