From help-request at octave dot org Fri Feb 4 20:42:05 2005 Subject: Re: octave-speed From: Paul Kienzle To: =?ISO-8859-1?Q?Rafael_Rodr=EDguez_Velilla?= Cc: help at octave dot org Date: Fri, 4 Feb 2005 21:45:50 -0500 Performance for vector operations is on par with matlab. Performance=20 for loops takes a 10x hit. You can develop the algorithm in Octave and code the bits that require=20= loops in C++. Look through the list archives for links to work on an octave compiler. - Paul On Feb 4, 2005, at 9:34 AM, Rafael Rodr=EDguez Velilla wrote: > Hello all, > > I'm new to octave and I'm planning to use it to simulate DSP=20 > algorithms. > > Speed is a matter of concern in my simulations. I have made a stupid=20= > test to compare octave's performance with Matlab's performance: > > m=3D1:100000; > n=3D8; > for i=3D1:99999 > n=3Dsqrt(sqrt(m(i)*m(i+1))*n); > endfor > n > > In the same computer it takes 10 seconds to run under linux-octave but=20= > less than a second under Matlab. > > I have compiled octave 2.1.64 with shared libraries with gcc3.2... > ... is there a way to improve its performance? > For example by compiling it with CPPFLAGS=3D-march=3Di686??? > > I have read that maybe using ATLAS improves the performance. > > Any hints on how to compile it to run faster? > > > > ------------------------------------------------------------- > Octave is freely available under the terms of the GNU GPL. > > Octave's home on the web: http://www.octave.org > How to fund new projects: http://www.octave.org/funding.html > Subscription information: http://www.octave.org/archive.html > ------------------------------------------------------------- > ------------------------------------------------------------- Octave is freely available under the terms of the GNU GPL. Octave's home on the web: http://www.octave.org How to fund new projects: http://www.octave.org/funding.html Subscription information: http://www.octave.org/archive.html -------------------------------------------------------------