Re: Wikipedia dump file processing shoot out

2009/12/11 Tyler Riddle <[email protected]>:
>>> Has anyone tried to identify WHY libxml/xslt running in perl is AT ALL
>>> slower than C? Obviously there is SOME going back and forth between perl and
>>> C there, but it kind of seems hideously more expensive than it should be.
>>>
>> There's an old thread about this [1]. Basically the performance penalty
>> comes from the XS interface: invoking an XS method is very expensive [2].
>> I've encounter this when I started writing Xacobeo [3] in which I was
>> displaying an XML document by traversing a DOM tree with XML::LibXML in pure
>> Perl.
>>
>
> I'm really curious why libxml's perl SAX interface performs so poorly
> compared to the others especially because libxml's SAX interface in C
> slightly outperforms expat:
>
>                           'name' => 'libxml.t',
>                           'percentage' => 100,
>
>                           'name' => 'expat.t',
>                           'percentage' => 112,
>
>                           'name' => 'XML-SAX-ExpatXS.t',
>                           'percentage' => 526,
>
>                           'name' => 'XML-Parser.t',
>                           'percentage' => 676,
>
>                           'name' => 'XML-SAX-ExpatXS_nocharjoin.t',
>                           'percentage' => 981,
>
>                           'name' => 'XML-LibXML-SAX.t',
>                           'percentage' => 1082,
>
> Actually now that I look at the data stacked up like that it makes
> perfect sense - it's not really that much slower than expat, it's
> XML::SAX::ExpatXS's character joining feature that is giving it the
> speed benefit. I have RT ticket #52368 open with XML-LibXML to add
> this exact feature but I have not heard anything back on it yet.

Tyler,

I've seen your RT ticket adding this feature to XML-LibXML and it sure
makes sense and I sure do plan to include your patch (or a revised
version of it perhaps) in the next release, so thanks a lot and sorry
for not responding to it right away. I usually answer invalid tickets
as they arrive, but the valid ones I tend to work on in batches when I
have more time.

-- Petr
_______________________________________________
Perl-XML mailing list
[email protected]
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs