Philippe Grosjean
2004-Apr-22 13:31 UTC
[R] New version of benchmark comparing R with other software
Hello,
Thanks to Douglas Bates, there is now a new benchmark suite (version 2.3)
which is compatible with R 1.9.0 and the recent Matrix library (0.8-1 or
above). You find it at http://www.sciviews.org/other/benchmark.htm. It
compares R 1.9.0 under Windows with:
S-PLUS 6.5, Matlab 6.0, O-Matrix 5.6, Octave 2.1.42, Scilab 2.7 and Ox 3.30.
In short, R in its version 1.9.0 and with the new Matrix library, is now one
of the fastest matrix calculation package among those tested on the computer
and system used (P IV, 1 Gb RAM and Windows XP pro).
Keeping its limitations in mind (only a few functions tested, and relatively
artificial situations that may or may not compare with real-world cases),
this benchmark suite could be useful for deciding which software to chose
for computing intensive matrix calculations, and also to compare the same
software (R) on different platforms, with one or several processors, and/or
with various optimized BLAS libraries.
Best,
Philippe Grosjean
.......................................................<??}))><....
) ) ) ) )
( ( ( ( ( Prof. Philippe Grosjean
\ ___ )
\/ECO\ ( Numerical Ecology of Aquatic Systems
/\___/ ) Mons-Hainaut University, Pentagone
/ ___ /( 8, Av. du Champ de Mars, 7000 Mons, Belgium
/NUM\/ )
\___/\ ( phone: + 32.65.37.34.97, fax: + 32.65.37.33.12
\ ) email: Philippe.Grosjean at umh.ac.be
) ) ) ) ) SciViews project coordinator (http://www.sciviews.org)
( ( ( ( (
...................................................................
Merci, Philippe.
Just a complement regarding Matlab 6.5 (R13).
R 1.9.0 is globally faster than Matlab 6.5 (tested on an AthlonXP):
Matlab R
Total time for all 15 tests (sec) 14.29 10.81
Overall mean (sec) 0.86 0.45
Matlab 6.5 has optimized the loops performance (sth called "Just In
Time"),
with spectacular results, codes with loops could be faster then vectorized
codes. Here's the result for the "loops" test:
Matlab R
Creation of a 220x220 Toeplitz matrix (loops) 0.01 0.50
Bravo to the R dev team, nice job !
Cheers
--
Fan
Philippe Grosjean wrote:> Hello,
>
> Thanks to Douglas Bates, there is now a new benchmark suite (version 2.3)
> which is compatible with R 1.9.0 and the recent Matrix library (0.8-1 or
> above). You find it at http://www.sciviews.org/other/benchmark.htm. It
> compares R 1.9.0 under Windows with:
> S-PLUS 6.5, Matlab 6.0, O-Matrix 5.6, Octave 2.1.42, Scilab 2.7 and Ox
3.30.
>
> In short, R in its version 1.9.0 and with the new Matrix library, is now
one
> of the fastest matrix calculation package among those tested on the
computer
> and system used (P IV, 1 Gb RAM and Windows XP pro).
>
> Keeping its limitations in mind (only a few functions tested, and
relatively
> artificial situations that may or may not compare with real-world cases),
> this benchmark suite could be useful for deciding which software to chose
> for computing intensive matrix calculations, and also to compare the same
> software (R) on different platforms, with one or several processors, and/or
> with various optimized BLAS libraries.
>
> Best,
>
> Philippe Grosjean
>
>
.......................................................<??}))><....
> ) ) ) ) )
> ( ( ( ( ( Prof. Philippe Grosjean
> \ ___ )
> \/ECO\ ( Numerical Ecology of Aquatic Systems
> /\___/ ) Mons-Hainaut University, Pentagone
> / ___ /( 8, Av. du Champ de Mars, 7000 Mons, Belgium
> /NUM\/ )
> \___/\ ( phone: + 32.65.37.34.97, fax: + 32.65.37.33.12
> \ ) email: Philippe.Grosjean at umh.ac.be
> ) ) ) ) ) SciViews project coordinator (http://www.sciviews.org)
> ( ( ( ( (
> ...................................................................
>
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide!
http://www.R-project.org/posting-guide.html
>