Hi Stephan,
I've updated my simple ruby xml benchmark which measures the time to open a 98k XML document and count one type of leaf element (466 entries) 100 times.https://svn.concord.org/svn/projects/trunk/common/ruby/xml_benchmarks/The summary is that for this one test using libxml in MRI and jdom_document_builder (a native Java library) in JRuby take just about the same time. This is similar to the last time I ran these benchmarks.
Nice. FYI - you know your script is in the libxml-ruby source tree in the script directory?
On my laptop (Thinkpad T60, Core 2 duo, 2GB memory):
user system total real
rexml 21.747000 0.000000 21.747000 ( 22.145000)
hpricot 4.633000 0.000000 4.633000 ( 4.761000)
libxml 0.468000 0.000000 0.468000 ( 0.463000)
Its impressive also how quickly Java is.
Aaraon - I added Nokogiri to the script, but it causes a general
protection fault. Maybe I'm doing something wrong (just checked code
into libxml-ruby trunk if you want to look)?
JRuby 1.1.5 has gotten much faster running pure Ruby code and now runs rexml about 30% faster than MRI. Hpricot is about 5 times slower than libxml and jdom_document_builder and runs slightly faster in JRuby.
Impressive. Charlie
smime.p7s
Description: S/MIME Cryptographic Signature
_______________________________________________ libxml-devel mailing list libxml-devel@rubyforge.org http://rubyforge.org/mailman/listinfo/libxml-devel
