[Haskell-cafe] Memory Leak - Artificial Neural Network
Shelby Moore
shelby at coolpage.com
Fri Nov 6 10:39:21 EST 2009
wren ng thornton wrote:
http://www.haskell.org/pipermail/haskell-cafe/2009-November/068728.html
>...use matrices/arrays rather than lists...
>...even just using an unboxed array for the matrices
> will make it much faster to traverse through...
Matrix multiplication is the sequential order inner or outer product of
the columns/rows of one matrix and rows/columns of 2nd matrix:
http://en.wikipedia.org/wiki/Matrix_multiplication#Relationship_with_the_inner_product_and_the_outer_product
Afaics, optimize matrix multiplication with lists. Random arrays incurs a
huge performance penality in Haskell, unless you revert to non-lazy
implementation, in which case you have voided multi-core speed ups for
very large arrays (aren't ANNs huge matrices?).
More information about the Haskell-Cafe
mailing list