sriram.durbha:
> it is a kind of folk lore that Haskell is not for numerically heavy code,
> do it in C and use FFI. Well has any one tried? and proved beyond doubt
I think that's actually an old meme. E.g.
http://hackage.haskell.org/cgi-bin/hackage-scripts/package/uvector
Is highly efficient, and yields much the same assembly as gcc finds.
Some details here,
http://cgi.cse.unsw.edu.au/~dons/blog/2008/06/04#fast-fusion
So time to update your expectations :)
> that FFI is the only way to go? some FFI may be end up being necessary
> but lower the level it is at, the better:) kind of like Data.ByteString
> although the actual packed string may be native, since there is a List
> interface, operations are in Haskell. but with hmatrix etc.. the
> operations are also native, we get an interface to those too. nothing
> wrong with using that!! just that trying to do it cleanly in Haskell, or
> another language might be a good thing.
You want some kind of unboxed matrix types ... perhaps data parallel
ones...
http://www.haskell.org/haskellwiki/GHC/Data_Parallel_Haskell
A bit new, but you can start playing with the vectorisation in ghc 6.10
next week. Some keen users already are.
-- Don
--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to "Bay Area Functional
Programmers"
To unsubscribe from this group, send email to [EMAIL PROTECTED]
For more options, visit this group at http://groups.google.com/group/bayfp?hl=en
-~----------~----~----~----~------~----~------~--~---
