Claus Reinke ha scritto:
Can I close this ticket as not being to do with uvector?
-- Don
You did notice the suggestion that performance of uvector and bytestring
could be improved drastically if compile-time fusion would be augmented
with runtime fusion?
The storablevector package implements Data.StorableVector.Lazy
Just as with Data.ByteString.Lazy, it contains a linked list of chunks.
I think that this can improve performances of my implementation, since
it is much more efficient to append elements at the end of the vector
(it avoids a lot of copying).
In my draft implementation of the the Netflix Prize in D language, I
used a similar implementation, base on:
http://mdounin.ru/hg/nginx-vendor-current/file/tip/src/core/ngx_list.h
Unfortunately, D garbage collector is really bad when there are a lot of
allocations, so I gave up.
Manlio Perillo
_______________________________________________
Haskell-Cafe mailing list
Haskell-Cafe@haskell.org
http://www.haskell.org/mailman/listinfo/haskell-cafe
