Hi!
> Sorry I was wrong,
> when doing it properly (caching everything that can be
> cached, and using the Translet), xsltc performs much better
> than xalan (twice faster in my examples), and even a bit
> better than Saxon.
Couldn't you show send me the code ?
The following was used in benchmarks:
// Create a transform factory instance.
TransformerFactory tfactory =
TransformerFactory.newInstance();
Templates translet = tfactory.newTemplates(new
StreamSource(sStylesheet));
Transformer transformer =
translet.newTransformer();
// Do an untimed run
StringWriter output=null;
for (int i=0; i<runs || i<1 ; i++) {
output=new StringWriter();
transformer.transform(new StreamSource(new
StringReader(document)),
new
StreamResult(output));
}
// Start test loop
long start=tick();
long totalsize=0;
for (int i=0; i<runs || i<1 ; i++) {
output=new StringWriter();
transformer.transform(new StreamSource(new
StringReader(document)),new StreamResult(output));
totalsize+=output.getBuffer().length();
}
long end=tick();
// If we are in timed mode report
if (runs>0) {
double size=totalsize/runs;
double
time=((double)(end-start))/(runs*1000);
double tput=(size/1024)/time;
local.println("<result>");
local.println("<size>" + size +
"</size>");
local.println("<time>" + time +
"</time>");
local.println("<tput>" + tput +
"</tput>");
local.println("</result>");
} else {
local.println(output.getBuffer());
}
>
> But, when using it as a normal transformer (Template instead
> of Translet) both of them (xalan and xsltc) give me the same
> results (xsltc even a bit worse).
>
> The question is (and sorry for my lack of knowledge): does
> Cocoon take advantage of the compilation features in xsltc?
>
> regards,
> Ruben
>
>
>
> |---------+---------------------------->
> | | |
> |---------+---------------------------->
>
> >-------------------------------------------------------------
> -------------------------------------------------|
> |
> |
>
> >-------------------------------------------------------------
> -------------------------------------------------|
>
>
>
>
> Hi Mikhail,
> I've done some very simple comparission tests (transforming
> reusing the
> templates)
> and xalan still performs a bit better than xsltc for me as
> well. If you are interested I can send you the results, but
> please, don't trust them, probably I'm missing something in
> xsltc configuration or something...
>
> btw, saxon was much faster than both of them.
>
> regards,
> Ruben
>
> -------------------------------------------------------------------
> Hi!
>
> > That would youi say about xslt benchmarks on
> http://www.sarvega.com/ ?
>
> > xsltc is rated even worse then
> xalanj itself...
>
> Still no reply. Comments anyone ? I wonder that was the cause
> of such perfomance of xsltc in those tests - tests
> themselves, or particular misfeatures of xsltc triggered by
> those tests ?
>
> My apologies if noone interested.
>
> Mikhail
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, email: [EMAIL PROTECTED]
>
>
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, email: [EMAIL PROTECTED]
>
>
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [EMAIL PROTECTED]
> For additional commands, email: [EMAIL PROTECTED]
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]
