Github user MechCoder commented on the pull request:
https://github.com/apache/spark/pull/5946#issuecomment-117989146
Oh just a second. I do get some improvements. I was benching the wrong
things in the previous comment.
There are averaged over 10 runs.
1. For a random sparse vector of length 500000, with 50000 values nd a
random DenseVector of the same length.
Timings: 0.06 s in master
A slightly less optimistic sparse vector of length 50000 and 500 values and
a random DenseVector of the same length
Timings: 0.0005s in master
Timings: 2.3937225341796876e-05 in master
With length 50000 and 5000 values.
Timings: 0.0058s in master
Timings: 7.47e-5 s in this branch
Squared distance
With length 50000 and 500 values.
Timings: 0.254s in master
Timings: 0.0003 s in this branch
With length 50000 and 5000 values.
Timings: 0.26862592697143556 in master.
Timings: 0.0008 s in this branch
With length 500000 and 50000 values
Timings in master: 2.352340269088745
Timings in this branch: 0.004776120185852051
Looks like we have a winner here. Do you want me to bench on anything more
specific?
---
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]
