[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Slowup in 2.1.54

From: Paul Kienzle
Subject: Re: Slowup in 2.1.54
Date: Wed, 18 Feb 2004 19:48:58 -0500

On Feb 18, 2004, at 7:14 PM, David Bateman wrote:

In any case this benchmark pretty much explains much of my recent
contributions. The stuff I did on sort and randn recently in
octave-forge, also the eigenvalue patch that doesn't calculate
eignevectors if you don't want them, and the FFT stuff. With these
changes and the fact the 2.1.42 didn't properly use LAPACK/ATLAS,
where Octave does now, I estimate the final total in this benchmark
for the CVS version of octave to have come down from 27.76 to about 15
or 16. Whether this reflects the speedup a real user will see is
more questionable however....

I certainly appreciate your work.  sort in particular is used
in many places it shouldn't be because it can avoid a
scripted for-loop. O(n log n) is faster than O(n) even on
large problems  if the constants are large enough.

So there is a direct benefit for things like hist and interp1.

Similarly, fft is used all over the place, so speedups here
affect much more than just the fft.

Fast random number generators are important in many
contexts, but maybe not so much in a scripted environment.
With a little massaging, though, we can have something
state of the art useful outside of octave.

Again, thanks!

Paul Kienzle

reply via email to

[Prev in Thread] Current Thread [Next in Thread]