comp.soft-sys.matlab Google Group

Re: Vectorization

guerra.remove.t...@physics.harvard.edu (Rodrigo) — Thu, 11 Mar 2010 22:57:04 UT

How did you get ndfun to work in 64 bits? I tried pointing mex to the blas and lapack libraries included in $MATLABROOT/bin/glnxa64 (I am using the 64bit linux version) and after fixing one or two of the comments it compiles. However, it segfaults whenever I run it.

Re: please help : ??? Index exceeds matrix dimensions.

clfei...@web.de (Hans Müller) — Thu, 11 Mar 2010 22:57:04 UT

Musste eben lachen, als ich das gelesen hatte. You can please the debugger kam etwas unklar bei mir an - du kannst dem Debugger eine Freude machen... mag er lieber Rosen oder Schokolade?
Grüsse

Re: How to achieve ultimate summing speed?

o.j.woodford...@cantab.net (Oliver Woodford) — Thu, 11 Mar 2010 22:56:04 UT

For this simple iteration over each column, with no dependencies between each one, the easiest way (IMHO) to parallelize this is with an OpenMP pragma before the first (outer) for loop. And it's completely platform independent.
Oliver

Re: making local variable global

nikitchmpub...@gmail.com (Max ) — Thu, 11 Mar 2010 22:52:04 UT

I know I can. What I'm worried about is that the message says this ability to turn local variable into global will be discontinued in the future. So I'm looking for an alternative way doing it.

Re: how to avoid a double for loop

ellieandrogerxy...@mindspring.com.invalid (Roger Stafford) — Thu, 11 Mar 2010 22:52:04 UT

You might try this. It uses only one call to 'unique'. The result is placed in B.
T = A.';
p = T(:,1)==1;
T(p,:) = 1-T(p,:); % Complement T rows that have 1 in first col.
[t,m,t] = unique(T,'rows'); % Reduce to unique T rows ('last')
B = A(:,sort(m)); % Reduce to corresponding cols. in A
Roger Stafford

Re: instead of For loop

mattjacrem...@thisieee.spam (Matt J ) — Thu, 11 Mar 2010 22:48:02 UT

Are you sure the sum isn't supposed to range over positive n-m?
Otherwise, you're inevitably going to get inf as the result (as Oleg did in his version of the calculation).

Re: How to achieve ultimate summing speed?

o.j.woodford...@cantab.net (Oliver Woodford) — Thu, 11 Mar 2010 22:43:04 UT

James, I'm surprised a speed demon like you is using mxCreateDoubleMatrix like that. Since it sets the matrix to zero first it drags the whole matrix through the cache once, before you even write to it. Since you then set every entry later you don't need to do that. See a discussion on the subject, and solution, here:

displaying "half" of a 3d plot

schreiber.pete...@gmail.com (Peter Schreiber) — Thu, 11 Mar 2010 22:37:04 UT

Hi guys,
I would like to display only half of a 3d plot to show some details of the functions.
More specificly, I want to make a diagonal cut through the 3d plot, and only display one half.
I could get it to work using plot3, but I really would need to represent a surface.
That's what I came up with so far:

Re: periodogram function

mul...@gmail.com (TideMan) — Thu, 11 Mar 2010 22:36:12 UT

Wayne
Are you sure about this?
I thought it transferred all the segments to the freq domain (using
FFT), squared (to get PSD), then took the mean.
If you do this in the time domain, as you suggest, you could
conceivably cancel the signal before it gets transformed.

Re: please help : ??? Index exceeds matrix dimensions.

gordan.sikic.rem...@this.inet.hr (Leclerc) — Thu, 11 Mar 2010 22:30:58 UT

Hi,
I do apologise,
I didn't want to offend anybody, I just tried to make a (obviously
unsuccessful) joke of entering discussion in the language I don't speak.
cheers,
Gordan

Re: Logging number of computations and memory ussage

sl...@mathworks.com (Steven Lord) — Thu, 11 Mar 2010 22:30:35 UT

That depends on how you define "effective". If you have a machine with very
little memory, but you can start the program running when you leave for the
night and get the results when you return in the morning, "effective" would
mean "memory efficient". On the other hand, if you need results RIGHT NOW

Re: instead of For loop

all...@tele.ntnu.no (Rune Allnor) — Thu, 11 Mar 2010 22:20:43 UT

You *will* have to use a for loop. You might hide
that fact behind the voodoo of 'vectorization',
but the for-loop will be there, nonetheless.
Apart from that, depending on the details that
are missing (the '-*' somebody already commented on)
you might be able to simplify the expression a bit,
possibly being able to exploit symmetries etc.

Re: instead of For loop

spama...@yahoo.com (Matt Fig) — Thu, 11 Mar 2010 22:19:05 UT

That is why I requested the actual code. If the OP *doesn't* get inf, then what was posted above in non-MATLAB notation is incorrect or something else is wrong. The OP claimed that the FOR loops were taking a long time, which seems to imply that there was some value being produced which was not inf. I guess we shall see ;-)

Re: instead of For loop

aclassyguy_with_a_k_not_...@hotmail.com (James Tursa) — Thu, 11 Mar 2010 22:19:05 UT

Well, one of the terms look like exp(-(0-999)*999) = exp(999*999) = inf.
So here is how to calculate the entire sum without any for loops:
inf
What are you doing?
James Tursa

Re: instead of For loop

oleg.komarovremove.t...@hotmail.it (Oleg Komarov) — Thu, 11 Mar 2010 22:09:05 UT

I interpret it as -1*(n-m). I implemente it as:
Out = sum(exp(reshape(-bsxfun(@times ,bsxfun(@minus, (0:999).',0:999),0:999),[],1)) );
which gives inf...
I'm sure it's some form of note matrix...but I'm not a matematician.
Oleg