[Ecls-list] string hashing

Juan Jose Garcia-Ripoll jjgarcia at users.sourceforge.net
Sun Jan 14 19:06:09 UTC 2007


2007/1/14, Andrew Schein <usenet at andrewschein.com>:
> I just ran the Marsden lisp benchmarks
> (http://www.chez.com/emarsden/downloads/) on ECL 0.9i to compare against
> SBCL 0.9.8.

Hi Andrew, how did you run the tests? Compiled or interpreted?

When I used to publish a benchmark page in ecls.sf.net I did each
benchmark twice: once compiling each function, one running them in the
interpreter. The reason is that some implementations compile
everything (namely SBCL) while others, using an interpreter, will
perform better/worse in different areas.

In any case these numbers are indeed useful to see in which areas ECL
must improve. Some of the tests are simple enough for other people to
investigate them or for instance run them in a version of ECL compiled
with profiling flags. Jusst random thoughts in case somebody finds
spare time :-)

Regards,

Juanjo

-- 
Dpto. de Fisica Teorica I, Fac. de CC Fisicas, Universidad Complutense,
Ciudad Universitaria s/n Madrid 2804 (Spain)
http://teorica.fis.ucm.es/~jjgarcia/




More information about the ecl-devel mailing list