Re: Very slow filter.cc

[John W. Eaton]

On 22-Jan-2005, Miroslaw Kwasniak <address@hidden> wrote:

| On Fri, Jan 21, 2005 at 03:20:52PM +0100, Miroslaw Kwasniak wrote:
| > On Wed, Jan 19, 2005 at 09:45:37AM +0100, David Bateman wrote:
| > > 
| > > t = cputime; y = filter(b,1,x); cputime - t
| > > 
| > > rather than using tic/toc and see if there is better consistency between 
| > > 2.0.17 and the 2.1.x versions...
| 
| Test code:
| 
|  OCTAVE_VERSION
|  x = randn(1e6,1);
|  b = randn(500,1);
|  t = cputime;
|  y = filter(b,1,x);
|  cputime - t
| 
|  
| > $ sh run.sh
| > OCTAVE_VERSION = 2.0.17
| > ans = 2.0600
| > OCTAVE_VERSION = 2.1.50
| > ans = 15.380
| > OCTAVE_VERSION = 2.1.57
| > ans = 15.690
| > OCTAVE_VERSION = 2.1.63
| > ans = 14.060
| > 
| > 2.1.63 is a little better because it's compiled for 686 - rest are debian
| > 386 packages.
| 
| 
| The same machine as above (Athlon 1600XP), fresh cvs source
| 
| OCTAVE_VERSION = 2.1.64
| ans = 4.1000
| 
| Hmmm, ...
| 
| Only important diff between 2.1.63/2.1.64 octave_config_info output:
| 
| -  CC_VERSION = 3.3.4 (Debian 1:3.3.4-13)
| +  CC_VERSION = 3.3.5 (Debian 1:3.3.5-5)
| -  CXX_VERSION = 3.3.4
| +  CXX_VERSION = 3.3.5

You can easily compare the current code and the code from 2.0.x and
see if you can make it faster.  If you don't care whether you can
interrupt the filter operation, then you can remove the OCTAVE_QUIT;
lines in the new code and it might run a little faster.  I don't think
that is a reasonable thing to do for most Octave users.

jwe



-------------------------------------------------------------
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
-------------------------------------------------------------