Thread Subject:

multiplying diag matrix with a reg matrix

I need to perform the product D*A where D is a diagonal matrix whose diagonal elements are non-zero and A a regular matrix with most of the elements non-zero. Is there is an operation or one-line statement that performs this taking into consideration that D is diagonal instead doing a literal matrix multiplication. This would save a lot of time for me.

regs
bala

"Balachander Narasimhan" <balxxxxxchand@whatever.com> wrote in message <gr8g5t$gti$1@fred.mathworks.com>...
> I need to perform the product D*A where D is a diagonal matrix whose diagonal elements are non-zero and A a regular matrix with most of the elements non-zero. Is there is an operation or one-line statement that performs this taking into consideration that D is diagonal instead doing a literal matrix multiplication. This would save a lot of time for me.
>
> regs
> bala

bsxfun(@times,diag(D),A)

OR

sparse(D)*A

Bruno

"Balachander Narasimhan" <balxxxxxchand@whatever.com> wrote in message <gr8g5t$gti$1@fred.mathworks.com>...
> I need to perform the product D*A where D is a diagonal matrix whose diagonal elements are non-zero and A a regular matrix with most of the elements non-zero. Is there is an operation or one-line statement that performs this taking into consideration that D is diagonal instead doing a literal matrix multiplication. This would save a lot of time for me.
>
> regs
> bala

  Turn D into a column vector and try the 'bsxfun' function with @times. It might be faster.

Roger Stafford

A = round(rand(1000)*10000);
D = diag(randperm(1000));

There are several ways to do this. Here are two shown in comparison.





N = 1000;
A = round(rand(N)*N^2);
D = diag(randperm(N));

tic
H = D*A;
toc

tic
C = bsxfun(@times,A,D(1:N+1:end).'); % Or use diag(D)
toc

tic
E = sparse(D)*A;
toc

all(H(:)==C(:))
all(H(:)==E(:))

"Matt Fig" <spamanon@yahoo.com> wrote in message <gr8hit$cto$1@fred.mathworks.com>...
> A = round(rand(1000)*10000);
> D = diag(randperm(1000));
>
> There are several ways to do this. Here are two shown in comparison.
>
>
>
>
>
> N = 1000;
> A = round(rand(N)*N^2);
> D = diag(randperm(N));
>
> tic
> H = D*A;
> toc
>
> tic
> C = bsxfun(@times,A,D(1:N+1:end).'); % Or use diag(D)
> toc
>
> tic
> E = sparse(D)*A;
> toc
>
> all(H(:)==C(:))
> all(H(:)==E(:))

Except that creating a dense matrix D then doing sparse(D) is slow. Better would be

d = randperm(N);
tic
E = sparse(1:n,1:n,d)*A;
toc

which never forms the dense D matrix, and is much faster than

E = sparse(D)*A;

"Tim Davis" <davis@cise.ufl.edu> wrote in message
> Except that creating a dense matrix D then doing sparse(D) is slow. Better would be
>
> d = randperm(N);
> tic
> E = sparse(1:n,1:n,d)*A;
> toc
>
> which never forms the dense D matrix, and is much faster than
>
> E = sparse(D)*A;



Very true Tiim, but the OP stated he already had the diagonal matrix. It turns out that my answer was the same as Bruno's (that's what I get for not hitting refresh on the browser before posting), so it seems we both took the problem as starting from that point.

"Matt Fig" <spamanon@yahoo.com> wrote in message <graki3$f5$1@fred.mathworks.com>...
> "Tim Davis" <davis@cise.ufl.edu> wrote in message
> > Except that creating a dense matrix D then doing sparse(D) is slow. Better would be
> >
> > d = randperm(N);
> > tic
> > E = sparse(1:n,1:n,d)*A;
> > toc
> >
> > which never forms the dense D matrix, and is much faster than
> >
> > E = sparse(D)*A;
>
>
>
> Very true Tiim, but the OP stated he already had the diagonal matrix. It turns out that my answer was the same as Bruno's (that's what I get for not hitting refresh on the browser before posting), so it seems we both took the problem as starting from that point.

True. I didn't mean to sound like I was critiquing your reply.

I just meant to suggest to the OP that it's not a good idea to from the dense version of the D matrix in the first place.

"Tim Davis" <davis@cise.ufl.edu> wrote in message
> True. I didn't mean to sound like I was critiquing your reply.
>
> I just meant to suggest to the OP that it's not a good idea to from the dense version of the D matrix in the first place.

O.k., but I don't mind a good critiquing now and then! : )

hi all, tnx for the replies, that solves my problem.

regs
bala

"Matt Fig" <spamanon@yahoo.com> wrote in message <grbp7r$j70$1@fred.mathworks.com>...
> "Tim Davis" <davis@cise.ufl.edu> wrote in message
> > True. I didn't mean to sound like I was critiquing your reply.
> >
> > I just meant to suggest to the OP that it's not a good idea to from the dense version of the D matrix in the first place.
>
> O.k., but I don't mind a good critiquing now and then! : )