help-octave
[Top][All Lists]
Advanced

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

Re: Vectorization quiz


From: Daniel Arteaga
Subject: Re: Vectorization quiz
Date: Thu, 28 Apr 2011 17:58:08 +0200
User-agent: Mozilla/5.0 (X11; U; Linux i686; ca-ES; rv:1.9.2.14) Gecko/20110223 Thunderbird/3.1.8

Al 21/04/11 16:37, En/na James Sherman Jr. ha escrit:
On Thu, Apr 21, 2011 at 4:49 AM, Daniel Arteaga
<address@hidden <mailto:address@hidden>> wrote:

    Hi,

    How could I vectorize the following piece of code?

    P = zeros(size(fInf));
    for i = 1:length(fInf)
            P(i) = sum( S2( fLin > fInf(i) & fLin < fSup(i) ) );
    endfor

    where

    fInf is a vector of length 1000
    fSup is a vector of length 1000
    P    is a vector of length 1000
    S2   is a vector of length 50000
    fLin is a vector of length 50000

    Thank you very much in advance,

    Daniel

    _______________________________________________
    Help-octave mailing list
    address@hidden <mailto:address@hidden>
    https://mailman.cae.wisc.edu/listinfo/help-octave


Well, I'm not saying that this is the best way, but you could try this:

n = length(fInf);
m = length(fLin);

fLin_repeated = repmat(fLin, [1, n]);
S2_repeated = repmat(S2, [1, n]);
fInf_repeated = repmat(fInf', [m, 1]);
fSup_repeated = repmat(fSup', [m, 1]);

tmp1 = (fLin_repeated > fInf_repeated) & (fLin_repeated < fSup_repeated);
P = sum( S2_repeated .* tmp1 )';

I particularly don't like the last line.  I feel like there is a
function to do the masking part instead of doing the multiplications,
but for the life of me I can't remember what it is.  The downside is
that you will have 4 different 50000 by 1000 matrices, which may or may
not be problematic.

Thank you very much for your answer. It comes down that I'm somewhat memory-limited, so it is not very convenient to have such big matrices. Maybe particioning S2 in smaller chunks as you suggest we could get a good compromise between memory consumption and speed.



reply via email to

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