[Haskell-cafe] Performance: MD5

[Brandon S. Allbery KF8NH]

On 2008 May 18, at 9:55, Andrew Coppin wrote:

> Brandon S. Allbery KF8NH wrote:
>>
>> Bulat is the naysayer of the group; according to him Haskell is a  
>> nice idea that will never actually work (but he uses it anyway, who  
>> knows how he rationalizes it to himself).
>>
>
> Bulat is apparently not alone in this belief. I seem to spend all my  
> time on other forums dealing with people who have exactly the same  
> opinion.

There's a difference.  When I started (GHC 6.4.2 was current), Bulat  
was spending his mailing list time denying the possibility of what DPH  
does now, and claiming that what GHC 6.8.2 does now was unlikely.

(Meanwhile, I can't say much for "oh, i didn't understand that code,  
but surely we should be able to do at least as good without  
performance hacks?" when the code you didn't understand is all  
performance hacks....  You really have no business trying to draw  
comparisons when you don't even know when you're comparing apples to  
aardvarks, let alone apples to oranges.)

Optimization is hard.  Don't like it?  Become a compiler researcher  
and find better ways to do it.  Note that C / procedural language  
optimization research has at least a 20-year head start on functional  
programming optimization, and that the problems are very different:   
the C world would love to be at the point where optimizing the C  
equivalent of "sum xs / length xs" is worth thinking about; they're  
still not really in a position to *detect* it unless the language is  
simple enough to make such reasoning relatively easy (e.g. FORTRAN).

-- 
brandon s. allbery [solaris,freebsd,perl,pugs,haskell] allbery at kf8nh.com
system administrator [openafs,heimdal,too many hats] allbery at ece.cmu.edu
electrical and computer engineering, carnegie mellon university    KF8NH