R help - Feb 2012 - using mclapply (multi core apply) to do matrix multiplication

Dear all,
I am trying to multiply three different matrices and each matrice is of size
16384,16384 the normal %*% multiplciation operator has not finished one day now.
As I am running a system with many cores (and it seems that R is using only one
of those) I would like to write fast a brief function that converts the typical
for loops of a matrix multiplication to a set of lapply sets (mclapply uses the
lapply syntax but it applies the work to many cores).

If my thinking is correct , in the sense that this will speed up things a lot, I
want you to help me covert the first matrix in rows the second in columns and
convert those in  a format that lapply would like to work with.


I would like to thank you in advance for your help

Regards
Alex

	[[alternative HTML version deleted]]

7-02-2012, 00:29 (-0800); Alaios escriu:
> Dear all, I am trying to multiply three different matrices and each
> matrice is of size 16384,16384 the normal %*% multiplciation
> operator has not finished one day now. As I am running a system with
> many cores (and it seems that R is using only one of those) I would
> like to write fast a brief function that converts the typical for
> loops of a matrix multiplication to a set of lapply sets (mclapply
> uses the lapply syntax but it applies the work to many cores).
>
> If my thinking is correct , in the sense that this will speed up
> things a lot, I want you to help me covert the first matrix in rows
> the second in columns and convert those in? a format that lapply
> would like to work with.
If I understand correctly, R uses a specialized library called BLAS to
do matrix multiplications. I doubt re-implementing the matrix
multiplication code at R-level would be any faster. What you can try
is replace BLAS with a multicore version of BLAS although it's not
easy if you have to compile it yourself.

Also, you may try to re-think the problem you're trying to solve.
Maybe there's a different approach that is less computation-intensive.

-- 
Cheers,
Ernest