xmlgraphics-fop-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jeremias Maerki <...@jeremias-maerki.ch>
Subject Re: Performance benchmark on new Intermediate Format
Date Mon, 16 Feb 2009 15:34:21 GMT
On 16.02.2009 15:56:33 Max Berger wrote:
> Jeremias,
> 
> Jeremias Maerki schrieb:
> > I've written a short report on the performance of the new Intermediate
> > Format. The numbers show that the goals have been met. It's also
> > interesting to look at from a general perspective. I'm sure there could
> > be a lot more that I could have written and shown but at some point I
> > had to warp it up. I hope I didn't leave out anything crucial.
> > 
> > http://people.apache.org/~jeremias/fop/benchmark-2009-02-13/
> 
> Looks really nice. Finally some numbers on FOPs performance, the tests
> you did seem to make sense. Thank you for this work! I think this should
> be merged as soon as possible.

Thanks. Having some hard numbers were one of the reasons why I invested
so much time in that. I often get the question: How fast is FOP? Now, we
have at least the beginning of an answer. Of course, more data is always
better. ;-)

> > My next step is to try to meet Vincent's high coding standards (although
> > that's probably a lost cause for me) and then to get the approval for
> > the merge into trunk.
> 
> I've just committed the following changes:
> - Fixed PMD to report only basic issues and important security issues.
> run "ant pmd" to get this report. This is something I'll look into
> changing in fop trunk after the merge.
> - added the PMD/CPD (Copy'n'paste detection) report, run "ant cpd" to
> generate it. Anything reported here should definitely be refactored.
> Again, I'll look into some of these things after your merge.

Cool. I didn't know it could do that. Pretty useful, and the result is
not half as bad as I expected after these discussions.

> PMD requires JDK 1.5 (only needed for report, and not for normal build
> or use, hope this is ok)

I don't see anything wrong with that.

> Hotspot does pretty well on inlining static functions, duplicate code
> could therefore go into static functions in utility classes if they
> would not make sense in the class hierachy.

Right, I just hope we don't go so far as to move every three duplicate
lines into their own utility method somewhere and thus make the code
unreadable.

> 
> > Have a good weekend,
> > Jeremias Maerki
> 
> hth
> 
> Max
> 
> -- 
> http://max.berger.name/
> OpenPGP ID: C93C5700 Fpr: AB6638CE472A499B3959 ADA2F989A2E5C93C5700
> 




Jeremias Maerki


Mime
View raw message