Re: [xml-dev] Rick Jelliffe's article on XSLT 1.0 performance

[Rick Jelliffe <rjelliffe@allette.com.au>]

Does anyone have any benchmarks on how much @as can improve performance, or any tips on which particular cases it helps.  I had kinda written it off as one of these theoretical improvements that never really pays off enough to be much use.

The reason is because we mooted it for Schematron XSLT3 Query Language Binding, and someone asked why not have it in the XSLT2 QLB too

Rick

On Sun, Feb 12, 2017 at 10:20 PM, Liam R. E. Quin <liam@w3.org> wrote:

On Sun, 2017-02-12 at 14:16 +1100, Rick Jelliffe wrote:
> Thanks Michael. And congrats on Saxon doing so well: Saxon HE being
> just as
> fast or faster than libxslt (ignoring JVM startup time) was one of
> the
> surprise results to me.

On my own Web site (www.fromoldbooks.org) in the few places where I use
a standalone Java query engine, I run it using "nailgun", which keeps a
pool of JVMs started and ready, runs the task on one of them and then
after sending back the result starts another JVM ready for next time.

I often do find Saxon faster than libxslt these days, even including
JVM startup time, and using some XSLT 2 techniques such as "as"
attributes on variables can help widen that gap.

Liam


--
Liam R. E. Quin <liam@w3.org>
The World Wide Web Consortium (W3C)