[Shootout-list] fannkuch benchmark for GHC

Isaac Gouy igouy2@yahoo.com
Thu, 6 Jan 2005 11:50:28 -0800 (PST)


> > Fannkuch is supposed to be a test of indexed-access to a
> > collection.

> I don't see that mentioned anywhere on the description page

What did you think it's about?
The paper mentions "manipulation of small integers and vectors of them"
The page does say "Indexed-access to tiny integer-sequence"


> I don't see that mentioned anywhere on the description page.  In
> fact, now that I look closer, I don't see any mention of whether this
> is supposed to be a "same way" or "same thing" benchmark. 

"same way" / "same thing" are so 2001 
Be careful what you ask for, you might get it ;-)

It does say "Each program ***should***"


> > Is this program and the Python program faster because it does fewer
> > accesses to the collection? (Better algorithm not faster access).
> 
> The paper linked from the description page mentions changing the
> algorithm in order to increase speed.  If that's not allowed, we
> should probably mention it.

It mentions *changing* the algorithm, it mentions a *better* algorithm,
(Iterative fibonacci, memoising fibonacci...) 

The page states  "Each program should"

It was a serious question - is that why they are faster?



		
__________________________________ 
Do you Yahoo!? 
The all-new My Yahoo! - Get yours free! 
http://my.yahoo.com