Good results here David.
You have to be very careful when doing micro-benchmarking. It is very
easy to see a couple of results here and make completely incorrect
conclusions. In other words it can present a very artificial world.
Now, the results here do not surprise me, i would expect the Java
version to outperform the CFML version by a large magnitude. This is
because we do not "compile" the CFML/CFC pages into Java byte code.
This has been well documented and discussed over the years. Other
engines, effectively convert to raw servlets and then run, which has
some advantages, but also has some disadvantages. For example, in our
tests, our technique of interpretation performs best at high scale,
using significantly less memory.
As for the Javascript side of things, we will look at resolving the last
remaining pieces (though i contend your choice of words "extremely
buggy" is somewhat misleading). The Rhino engine, which is what runs
the Javascript under the covers is something that is continually
evolving from the Mozilla team and will be updated in subsequent Java
releases.
Thanks for the input here. Great stuff.
--
aw2.0 ltd ::: http://about.me/aw2
David Mulder wrote:
Last week I spent half an hour comparing /coldfusion /vs /coldfusion
cfscript/ vs/java cfscript/ vs /javascript cfscript/ and I was kind of
surprised by the result, so I thought I would share it (+I saw on
another thread some questions about cfscript javascript speed):
* CFscript language=javascript is quite extremely buggy: issue #416
and #417
* CFscript is about the same speed as coldfusion
* *CFscript language=java is 100 (!!!) times as fast as coldfusion*
in calculating pi*.*
Especially the third point I would love to hear *why *the difference
is *that* big.
--
online documentation: http://openbd.org/manual/
google+ hints/tips: https://plus.google.com/115990347459711259462
http://groups.google.com/group/openbd?hl=en
Join us @ http://www.OpenCFsummit.org/ Dallas, Feb 2012
