Hi Folks, I've been collating timing info on the loop tic; for i=1:1e7, end; toc posted on the help-octave and sus-linux-e lists (I'm using SuSE-7.2). So far, it comes out as follows. The timings in square brackets are converted by scaling for CPU MHz to 733MHz, and the results are in order of this. CPU/MHz Oct ver time Linux ver ----------- ------- ----------- --------- PIII/600 ?(U/C) 17s [14s] (S)SuSE-7.2 PIII/1000 2.1.34 12.5s [17s] SuSE-7.2 PII/350 2.1.34 44s [21s] ? Athlon/1100 2.1.33 16s [24s] ? K7/700 2.1.28 27s [26s] ? PIII/800 ? <27s [<29] ? K7/1200 ? 20s [33s] ? Me *PIII/733 2.0.16 45s [45s] (S)SuSE-7.2 *PIII/733 2.0.16 50s [50s] (S)SuSE-7.2 The two cases marked with * are using the 2.0.16 binaries off the SuSE CDs, compiled for i386 architecture and therefore not optimised for other CPUs. Cases marked with (S) have been given on the SuSE list. The cases where other (later) versions of octave are given are probably user-compiled. The case with "U/C" is definitely user-compiled (according to the writer). For the other "?" cases no info was given. There is a pretty clear implication here: if you make serious use of octave for intensive computation, you should be able to speed it up by a factor of 2, if not 3, by compiling your own version from source; and it is probably advantageus to get a later version of octave as well. Best wishes to all, Ted. -------------------------------------------------------------------- E-Mail: (Ted Harding) <Ted.Harding@nessie.mcc.ac.uk> Fax-to-email: +44 (0)870 167 1972 Date: 11-Aug-01 Time: 12:16:00 ------------------------------ XFMail ------------------------------