Effacer les filtres
Effacer les filtres

Using Arrays inside arrayfun()

4 vues (au cours des 30 derniers jours)
Maxime Rivard
Maxime Rivard le 23 Déc 2017
I'm using gpuArrays with arrayfun to speed up calculations, and I would need to pass an array inside the call, as such
[Zn] = arrayfun( @iterFuncDelta,Xo,Yo,orbit)
function [Zn] = iterFuncDelta(xdelta0,ydelta0,refOrbit)
maxIterations=2^15;
delta=complex(xdelta0,ydelta0);
delta0=delta;
n=2;
Zn=refOrbit(1,1)+delta;
while ( n <= maxIterations )
Zn=refOrbit(n,1)+delta;
delta=2*refOrbit(n,1)*delta+delta^2+delta0;
n=n+1;
end
end
How could I pass the array 'orbit' so that it is not treated as a single point (like Xo and Yo) ?
Thanks.

Réponse acceptée

Matt J
Matt J le 23 Déc 2017
Modifié(e) : Matt J le 23 Déc 2017
According to this example, it should be possible with a nested function. Not sure how fast it's likely to be, though.
function Outer(Xo,Yo,orbit) %EDIT
refOrbit=orbit;
function [Zn] = iterFuncDelta(xdelta0,ydelta0)
maxIterations=2^15;
delta=complex(xdelta0,ydelta0);
delta0=delta;
n=2;
Zn=refOrbit(1,1)+delta;
while ( n <= maxIterations )
Zn=refOrbit(n,1)+delta;
delta=2*refOrbit(n,1)*delta+delta^2+delta0;
n=n+1;
end
end
[Zn] = arrayfun( @iterFuncDelta,Xo,Yo);
end
  2 commentaires
Maxime Rivard
Maxime Rivard le 23 Déc 2017
But in that case, how does the arrayfun() call get access to the arrays Xo and Yo?
Matt J
Matt J le 23 Déc 2017
Modifié(e) : Matt J le 23 Déc 2017
You would pass that to Outer() as well (see edit). I of course can't know what other variables you actually have in your workspace, nor how they got there. The inclusion of Outer() was just an example to show the nested context needed for iterFuncDelta.

Connectez-vous pour commenter.

Plus de réponses (2)

Walter Roberson
Walter Roberson le 23 Déc 2017
  3 commentaires
Walter Roberson
Walter Roberson le 23 Déc 2017
[Zn] = arrayfun( @(x,y) iterFuncDelta(x,y,orbit), Xo, Yo )
Matt J
Matt J le 24 Déc 2017
I don't think this will work on the GPU.

Connectez-vous pour commenter.


Joss Knight
Joss Knight le 26 Déc 2017
You can't do any array or matrix operations in a GPU arrayfun kernel. You can access the contents of an array that is present as an up-level variable such as has been demonstrated in the comments. You can then loop over the contents of the array one element at a time, to emulate array operations such as a dot product. This blog article gives an example of doing this for a simple 3-vector.

Catégories

En savoir plus sur GPU Computing dans Help Center et File Exchange

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by