Stefan O'Rear wrote:
On Sun, Jul 08, 2007 at 12:10:04PM +0100, Andrew Coppin wrote:
(Realistically though. My program takes a [Word8] and turns it into a [Bool] before running a parser over it. The GHC optimiser doesn't really stand a hope in hell of optimising that into a program that reads a machine word into a CPU register and starts playing with bit flips on it...)
Actually, if you're very lucky (fusion is just as hard in Haskell as it is in real life), it *does*. It seems to fit nicely into the stream-fusion framework.
Ooo... really? That's pretty impressive...(!) Is there a way I can check? ;-) More usefully, can I do stuff to my code to make myself more "lucky"? (Love the comment about RL BTW!)
PS. Are those zlib libraries actually written in Haskell? Or are they a thin layer on top of a C library?
Yup, they wrap C's zlib.
Thought so. Comparing native Haskell to a heavily optimised C library would surely be just like comparing native Haskell to a compiled C binary...
PPS. Does GHC make use of MMX, SSE, et al?
No (in spirit - the native code generator uses 1-element SSE operations for floating point because it's easier to optimize than "FPU" code).
Does GHC actually do anything that could usefully use these primitives? (I'm guessing "no"...)