Effacer les filtres
Effacer les filtres

How can i classified the variable inside of the parfor loop?

3 vues (au cours des 30 derniers jours)
PRIYANGA
PRIYANGA le 28 Jan 2013
Sir, This is my PARFOR loop coding. I can't classified some variables in PARFOR LOOP.
parfor i1=1:1:nrr1
disp(i1);
if (varran(i1) < threshold)
labe(i1)=0;
mvqcoder(i1)=[meanranr(i1)];
mvqcodeg(i1)=[meanrang(i1)];
mvqcodeb(i1)=[meanranb(i1)];
% e=e+1;
else
labe(i1)=1;
[domr]=domsearch(rang1r(i1,:),dompoolr,gsize,ci1r(i1),ci2r(i1),ci3r(i1),msr(i1),mdr,cd1r,cd2r,cd3r,mscr);
[domg]=domsearch(rang1g(i1,:),dompoolg,gsize,ci1g(i1),ci2g(i1),ci3g(i1),msg(i1),mdg,cd1g,cd2g,cd3g,mscg);
[domb]=domsearch(rang1b(i1,:),dompoolb,gsize,ci1b(i1),ci2b(i1),ci3b(i1),msb(i1),mdb,cd1b,cd2b,cd3b,mscb);
dom=[domr,domg,domb];
[rz cz]=size(dom);
meanrar=mean(rang1r(i1,:));
meanrag=mean(rang1g(i1,:));
meanrab=mean(rang1b(i1,:));
dompolr=iCalc(dompoolr,dom,cz);
dompolg=iCalc(dompoolg,dom,cz);
dompolb=iCalc(dompoolb,dom,cz);
[isomer alphr meanrr dom1r errr]=affwerrqt(rang1r(i1,:),dompolr(1:cz,:),gsize);
[isomeg alphg meanrg dom1g errg]=affwerrqt(rang1g(i1,:),dompolg(1:cz,:),gsize);
[isomeb alphb meanrb dom1b errb]=affwerrqt(rang1b(i1,:),dompolb(1:cz,:),gsize);
cdom1r(i1)=dom(dom1r);
isome1r(i1)=isomer;
alph1r(i1)=alphr;
meanr1r(i1)=meanrar;
cdom1g(i1)=dom(dom1g);
isome1g(i1)=isomeg;
alph1g(i1)=alphg;
meanr1g(i1)=meanrag;
cdom1b(i1)=dom(dom1b);
isome1b(i1)=isomeb;
alph1b(i1)=alphb;
meanr1b(i1)=meanrab;
vqcode3r(i1,:)=[isome1r(i1),alph1r(i1),meanr1r(i1),cdom1r(i1)];
vqcode3g(i1,:)=[isome1g(i1),alph1g(i1),meanr1g(i1),cdom1g(i1)];
vqcode3b(i1,:)=[isome1b(i1),alph1b(i1),meanr1b(i1),cdom1b(i1)];
% c=c+1;
end
% w1=w1+1;
end
I got Error result
The temporary variable meanrar is used after the PARFOR loop but its value is non deterministic.
The temporary variable meanrag is used after the PARFOR loop but its value is non deterministic.
The temporary variable meanrab is used after the PARFOR loop but its value is non deterministic.
then,
The PARFOR loop can't run due to the way variable dompolr is used.
The PARFOR loop can't run due to the way variable dompolg is used.
The PARFOR loop can't run due to the way variable dompolb is used.
Please any one help this problem for parfor loop.
thank you
  4 commentaires
PRIYANGA
PRIYANGA le 30 Jan 2013
Sir give a solution for this problem sir
Randy Souza
Randy Souza le 12 Fév 2013
I have restored the original text of this question.
PRIYANGA, this question has a clear subject and an accepted answer, so it may be valuable to someone else in the future. If you have a good reason why it should be removed from MATLAB Answers, please flag the question, explain why it should be deleted, and an administrator or high-reputation contributor will consider deleting the question. Please do not simply edit your question away.

Connectez-vous pour commenter.

Réponse acceptée

Edric Ellis
Edric Ellis le 28 Jan 2013
The problem is that you're building e.g. dompolr in a loop which the PARFOR machinery cannot understand. Your code currently looks a bit like this:
parfor i1 = 1:N
...
for z = 1:cz
dompolr(z,:) = dompoolr(dom(z),:);
end
...
end
I think it would fix things and simplify your code if you had instead
parfor i1 = 1:N
...
dompolr = iCalc(dompoolr,dom,cz);
...
end
...
function val = iCalc( dompoolx, dom, cz )
for z = 1:cz
val(z,:) = dompoolx(dom(z),:);
end
  3 commentaires
PRIYANGA
PRIYANGA le 30 Jan 2013
Sir give a solution for this problem sir
Edric Ellis
Edric Ellis le 1 Fév 2013
Any variables that are completely written to within the PARFOR loop are 'temporary', and therefore you cannot rely on the value after the loop has completed. For example
parfor ii = 1:3
x = ii;
end
disp(x)
In this case, the value of 'x' cannot be known because it depends on the order of evaluation of the PARFOR loop (which is not deterministic). So, you must not use temporary variables after the PARFOR loop - instead, all outputs must either be sliced (i.e. indexed using the loop variable), or reductions, like
x = 0;
parfor ii=1:3
x = x + ii;
end
disp(x);

Connectez-vous pour commenter.

Plus de réponses (0)

Catégories

En savoir plus sur Parallel Computing Fundamentals 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