[
https://issues.apache.org/jira/browse/RNG-75?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16854632#comment-16854632
]
Alex D Herbert commented on RNG-75:
-----------------------------------
{quote}Performance is a non-goal for this class.
{quote}
That was the impression I was working under. It was not clear from my post that
I was pointing out that the main cost of constructing a SplitMix generator is
the SeedFactory creating a new long. I agree that the output from the
SeedFactory is good.
I am currently running a Well44497b generator xor'd with the output of System
getIdentityHashCode through the stress test application. A quick checks shows
it passes SmallCrush (there was 1 fail from 5 runs for a plain Well44497b) and
is about 33% slower to run the test. This shows that the Java-to-C bridge is
the main bottle neck in the stress test application. Anyway it should be done
by tomorrow and I expect the output to pass. I do not suggest altering the
SeedFactory. The performance test was to gauge how to add new *create...Fast*
methods to the SeedFactory.
I have another table that shows the relative performance of using the
RandomSource.create method against using the *new* keyword to create the
generator, with a seed dynamically created by the SeedFactory and with a
pre-built seed. The pre-built seed case shows that reflection is a lot slower
when the seeding routine is simple. However I would assign the later as a
minority case. Building a generator in the majority case would require building
a new seed. I'll post this table later (as I do not have it on my current
workstation).
> Improve the speed of the RandomSource create method.
> ----------------------------------------------------
>
> Key: RNG-75
> URL: https://issues.apache.org/jira/browse/RNG-75
> Project: Commons RNG
> Issue Type: Improvement
> Components: simple
> Affects Versions: 1.3
> Reporter: Alex D Herbert
> Assignee: Alex D Herbert
> Priority: Minor
> Fix For: 1.3
>
> Attachments: large.jpg, long.jpg, small.jpg
>
>
> Update the {{o.a.c.rng.simple.internal}} package to improve the construction
> speed of random generators.
> Areas identified by the construction benchmark
> [RNG-72|https://issues.apache.org/jira/projects/RNG/issues/RNG-72] include:
> * Update the {{RandomSourceInternal}} to know the desired size for the native
> seed
> * Update the {{SeedFactory}} for faster {{byte[]}} conversions
> * Remove the use of reflection for fast seeding generators
> It is intended that all changes made are non-destructive to the quality of
> any generated seed.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
