[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: Assembly-language XML parser

• From: Rob Cameron <cameron@cs.sfu.ca>
• To: rjelliffe@allette.com.au
• Date: Sat, 9 Oct 2010 10:10:24 -0700 (PDT)

Thanks for the note, Rick.

But on the topic of points for bravery, we're now trying
to avoid that, by hiding the low-level parallel bit
stream and SIMD programming details using compiler
technology.   Our goal is to broaden the applicability
of the techniques well beyond XML.

This approach is the crux of our Parabix 2 effort.  At
Balisage '09 we presented the basic bistream equations
and concepts.   Subsequently, we have developed a
high-level Python prototyping technique and our first
prototype compiler, completed this summer.   

Of course, this is still a research effort and the XML 
community may be more interested in practical parsers.
On that note, we have now switched emphasis from development
of a standalone Parabix parser to development of
accelerated versions of existing tools (Xerces, Saxon, etc.).

To check out our research code and xmlwf sample application
(for DTDless XML), use the following steps (Linux).

(1) svn co http://parabix.costar.sfu.ca/svn/proto
(2) Change to the parabix2 prototype directory and compile
    to generate C code.
cd proto/parabix2
make
(3) Change to the src subdirectory and compile the C code.
cd src
make

You'll need an SSE 4.1 machine to run the xmlwf application.

(4) Test the application against the W3C XML conformance
suite (DTDless subset).
make test



Robert D. Cameron, Ph.D
Professor of Computing Science
Associate Dean of Applied Sciences
Simon Fraser University



----- Original Message -----
> > It's x86 assembler, and to fit my memories of assembler it would
> > have to
> > be 6502, but it's certainly more useful this way on modern
> > platforms:
> >
> > http://tibleiz.net/asm-xml/
> >
> > (via
> > <http://radar.oreilly.com/2010/10/four-short-links-8-october-201.html>)
> >
> > I haven't tried it, but they definitely get points for bravery.
> 
> Don't forget Cameron's parser*, which uses C/C++ intrinsic functions
> (which are just SIMD assembler instructions presented as C functions.)
> I'd
> count that as assembler.
> 
> Cheers
> Rick Jelliffe
> 
> *http://www.balisage.net/Proceedings/vol4/author-pkg/Cameron01/BalisageVol4-Cameron01.html
> 
> _______________________________________________________________________
> 
> XML-DEV is a publicly archived, unmoderated list hosted by OASIS
> to support XML implementation and development. To minimize
> spam in the archives, you must subscribe before posting.
> 
> [Un]Subscribe/change address: http://www.oasis-open.org/mlmanage/
> Or unsubscribe: xml-dev-unsubscribe@lists.xml.org
> subscribe: xml-dev-subscribe@lists.xml.org
> List archive: http://lists.xml.org/archives/xml-dev/
> List Guidelines: http://www.oasis-open.org/maillists/guidelines.php

• Follow-Ups:
  • Re: Assembly-language XML parser
    • From: Liam R E Quin <liam@w3.org>

• References:
  • Re: Assembly-language XML parser
    • From: rjelliffe@allette.com.au