[Haskell-cafe] Profiling help (Warning: Euler spoilers)

Louis Wasserman wasserman.louis at gmail.com
Thu Mar 4 10:07:51 EST 2010

Actually, looking back, I'm not sure mapM is even the right choice.  I think
foldM would suffice.  All we're doing is finding the association pair with
the minimum key, no?  In this case, foldM would do everything we need
to...and State.Strict would be pretty good at that.

Louis Wasserman
wasserman.louis at gmail.com

On Thu, Mar 4, 2010 at 8:32 AM, Daniel Fischer <daniel.is.fischer at web.de>wrote:

> Am Donnerstag 04 März 2010 04:20:05 schrieb Louis Wasserman:
> > James,
> >
> > Which version of Control.Monad.State are you importing?
> >
> > If you're just importing vanilla Control.Monad.State, that's actually
> > sending you to Control.Monad.State.Lazy.
> >
> > Using Control.Monad.State.Strict might solve your problems, srsly.
> > Laziness can result in epically failing memory leaks.
> That was my first thought too, but for this code, Control.Monad.State.Lazy
> is the better choice. Control.Monad.State.Strict doesn't play too well with
> sequence (and hence mapM). Since James consumes the list of chain lengths
> as it is produced (key = maximum keys), no large thunks can build up. With
> the strict State, he can't start to look for the maximum until all chain
> lengths have been computed, building a big fat thunk for the list.
> >
> > Louis Wasserman
> > wasserman.louis at gmail.com
> > http://profiles.google.com/wasserman.louis
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.haskell.org/pipermail/haskell-cafe/attachments/20100304/870b6d58/attachment.html

More information about the Haskell-Cafe mailing list