--- Holger_Fl�rke <[EMAIL PROTECTED]> wrote:
> Hi there,
> 
>    in a german computer magazin (Heiko W. Rupp,
> "Geschwindigkeitsrausch - 
> Leistungsfaehigkeit von XSLT-Prozessoren", iX 7/03)
> was an article about XSLT 
> benchmarking. The author has referenced two
> benchmarks: XSLTMark and XSLTBench. I had 
> a *short* look at Sarvegas XSLTBench
> (www.sarvega.com) and found XalanC and XalanJ 
> are worse than the average and are beaten by eg
> saxon and msxml. There are only one 
> test, where XalanC is faster than saxon.
> 
>    Anybody had a closer look at theese benchmarks?

I had a VERY close look at the benchmarks :-)
and I find Sarvega's XSLTBench much closer to
real life reality than XSLTMark;
for more info go to
http://www.ambrosoft.com/benchmarks.htm
BTW, Sarvega's results for XSLTC are wrong:
it finishes the race as one of the slowest processors
because stylesheets are recompiled in each iteration;
I have spoken to them about it.

> Especially: Why should XalanC be 
> slower than most of the Java processors?

Why not?  Java these days is not any more the synonym
of "slow." Algorithmic details make all the difference
C/C++ by themselves are no guarantee of speed.

> Are there
> any other "accepted" benchmarks?

Not to my knowledge.

> HolgeR

--Jacek

__________________________________
Do you Yahoo!?
SBC Yahoo! DSL - Now only $29.95 per month!
http://sbc.yahoo.com

Reply via email to