This is the mail archive of the
gsl-discuss@sourceware.org
mailing list for the GSL project.
check for 1.11.90 / timings for simplex2
- From: Ivo Alxneit-Kamber <ivo dot alxneit at psi dot ch>
- To: gsl-discuss at sourceware dot org
- Date: Fri, 12 Dec 2008 13:18:49 +0100
- Subject: check for 1.11.90 / timings for simplex2
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
hi everybody
no problems with compile and check on my system
uname -a
Linux pc6411 2.6.18-92.1.18.el5 #1 SMP
Wed Nov 12 06:49:06 EST 2008 i686 i686 i386 GNU/Linux
gcc -v
Using built-in specs.
Target: i386-redhat-linux
Configured with: ../configure --prefix=/usr --mandir=/usr/share/man
--infodir=/usr/share/info --enable-shared --enable-threads=posix
--enable-checking=release --with-system-zlib --enable-__cxa_atexit
--disable-libunwind-exceptions --enable-libgcj-multifile
--enable-languages=c,c++,objc,obj-c++,java,fortran,ada
--enable-java-awt=gtk --disable-dssi --enable-plugin
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-1.4.2.0/jre
--with-cpu=generic --host=i386-redhat-linux
Thread model: posix
gcc version 4.1.2 20070626 (Red Hat 4.1.2-14)
i did some timing checks with my application where i minimize
3 parameters and the evaluation of the target function is rather
costly. part of the function evaluation is done 'externally'
(in a 'system()' call) so you'll have to compare 'real' times below.
four different initial vertices are tested (each once) on a pretty
much idle system. reported are the values of the tree fitted
parameters, something comparable to chi2, and the size of the simplex.
what i find:
- - simplex2 find the same minimum as simplex
- - simplex2 reports slightly different values for the size of the
simplex. and because of this
- - simplex2 tends to take a few more iterations than simplex and is thus
slightly slower
- - in the case where both algorithms use the same number of iterations
(case 2) there is no significant speed difference.
does this make sense? brian, did you run some tests for fits with
'a lot' of parameters? thats where simplex2 should shine as far as
i understand.
brian could you please reconsider the inclusion of the alternate
procedure to set the initial simplex (SimplexImproved) for 1.12.
I never had any success with the original version in my applications.
including it would also maintain binary compatibility with existing code
as the new initializer gets only triggered if you specify a stepsize of
zero for all but the first parameter, a situation very unlikely to occur
otherwise.
timings
1)
simplex2 real 0m44.091s
simplex2 user 0m36.966s
simplex2 sys 0m7.051s
3.914659e-01 8.625244e-04 -8.103742e-02 9.673689e-11 9.678518e-05
# converged after 99 iterations
simplex real 0m44.918s
simplex user 0m37.789s
simplex sys 0m7.121s
3.914659e-01 8.625244e-04 -8.103742e-02 9.673689e-11 9.614379e-05
# converged after 103 iterations
2)
simplex2 real 0m57.502s
simplex2 user 0m48.285s
simplex2 sys 0m9.181s
3.914666e-01 8.626529e-04 -8.103258e-02 9.673689e-11 8.612845e-05
# converged after 129 iterations
simplex real 0m57.414s
simplex user 0m48.286s
simplex sys 0m9.113s
3.914666e-01 8.626529e-04 -8.103258e-02 9.673689e-11 7.646792e-05
# converged after 129 iterations
3)
simplex2 real 0m46.427s
simplex2 user 0m39.078s
simplex2 sys 0m7.330s
3.914675e-01 8.626788e-04 -8.098094e-02 9.673689e-11 8.378101e-05
# converged after 103 iterations
simplex real 0m46.946s
simplex user 0m39.554s
3.914675e-01 8.626788e-04 -8.098094e-02 9.673689e-11 7.898101e-05
# converged after 105 iterations
simplex sys 0m7.339s
4)
simplex2 real 0m38.351s
simplex2 user 0m32.287s
simplex2 sys 0m5.949s
3.914675e-01 8.625820e-04 -8.096469e-02 9.673689e-11 8.422464e-05
# converged after 85 iterations
simplex real 0m38.826s
simplex user 0m32.670s
simplex sys 0m6.052s
3.914675e-01 8.625820e-04 -8.096469e-02 9.673689e-11 8.318496e-05
# converged after 87 iterations
regards
- --
Dr. Ivo Alxneit
Laboratory for Solar Technology phone: +41 56 310 4092
Paul Scherrer Institute fax: +41 56 310 2688
CH-5232 Villigen http://solar.web.psi.ch
Switzerland gnupg key: 0x515E30C7
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Red Hat - http://enigmail.mozdev.org
iD8DBQFJQlaoAd7CE1FeMMcRAlEiAJ9gpb3Ocucmam4NmEuYUDPS5FxYzACeI458
oi5TL/BOTWTRnJUHmTNiynQ=
=sNkF
-----END PGP SIGNATURE-----