Github user mengxr commented on the pull request:
https://github.com/apache/spark/pull/1093#issuecomment-73376230
Does `FastMath` give significant performance improvement over `math`? I
think this needs some performance testing, given that there are other overheads
involved in the computation. If we don't see significant gain, maybe it is not
worthing using it. People might have different versions of commons-math3 on the
classpath, so we should try to use a minimal subset of its functions. From its
doc
(http://commons.apache.org/proper/commons-math/apidocs/org/apache/commons/math3/util/FastMath.html),
many functions are marked as "since 3.4". Another issue is the optimization
used in FastMath. For example, this is the `log` implementation:
https://github.com/apache/commons-math/blob/master/src/main/java/org/apache/commons/math3/util/FastMath.java#L1141
It might be faster than JVM's implementation. But if anything
(accuracy/performance) goes wrong there, it will be extremely hard for us the
trace the problem.
About the API, is it okay to put `kernelDensity` as a method under
`Statistics` and hide the implementation?
---
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]
