Github user debasish83 commented on the pull request:
https://github.com/apache/spark/pull/5005#issuecomment-81443187
I also printed how much time is taken in inner solve in Breeze NNLS and how
much time is in total solve (iterator pattern + other stuff). Looks like there
is some overhead:
15/03/15 22:15:54 INFO ALS: inner solveTime 92.635 ms
15/03/15 22:15:54 INFO ALS: solveTime 172.768 ms
15/03/15 22:15:54 INFO ALS: inner solveTime 92.032 ms
15/03/15 22:15:54 INFO ALS: solveTime 171.95 ms
15/03/15 22:15:54 INFO ALS: inner solveTime 41.084 ms
15/03/15 22:15:54 INFO ALS: solveTime 67.539 ms
15/03/15 22:15:54 INFO ALS: inner solveTime 43.021 ms
15/03/15 22:15:54 INFO ALS: solveTime 70.463 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 26.012 ms
15/03/15 22:15:55 INFO ALS: solveTime 30.391 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 26.347 ms
15/03/15 22:15:55 INFO ALS: solveTime 30.623 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 54.749 ms
15/03/15 22:15:55 INFO ALS: solveTime 60.712 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 52.006 ms
15/03/15 22:15:55 INFO ALS: solveTime 58.623 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 25.416 ms
15/03/15 22:15:55 INFO ALS: solveTime 29.712 ms
15/03/15 22:15:55 INFO ALS: inner solveTime 25.556 ms
15/03/15 22:15:55 INFO ALS: solveTime 29.974 ms
But interestingly inner solveTime is still ~5X of mllib NNLS which does not
make sense...I will take a closer look tomorrow...
---
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]
