Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]


Groups > comp.lang.haskell > #172

Re: arrays instead of lists

From Paul Rubin <no.email@nospam.invalid>
Newsgroups comp.lang.haskell
Subject Re: arrays instead of lists
References <4119f978-a247-4efa-93d1-af6d6ddcb5d9@b9g2000pbo.googlegroups.com>
Date 2012-02-20 01:17 -0800
Message-ID <7xr4xp7tkc.fsf@ruckus.brouhaha.com> (permalink)
Organization Nightsong/Fort GNOX

Show all headers | View raw


Sean McIlroy <namenobodywants@gmail.com> writes:
> how much more efficient is it to use (one-dimensional) arrays instead
> of (long) lists? could it conceiveably make the difference between a
> script that overflows the stack and one that doesn't? thanks if you
> can help

Stack overflow is usually caused by space leaks due to lazy evaluation
letting unforced thunks pile up.  It takes a while to develop a sense of
where this is happening and put in annotations to prevent it.  One easy
measure is to compile with ghc -O2 and see if that helps (it does more
automatic analysis of this issue).  

Once you have the space leaks out of the way, you shouldn't get stack
overflow unless the program uses an awful lot of memory.  Arrays can be
faster than lists or intmaps, but it's best to think of them as a
low-level optimization to pursue after you've gotten algorithmic issues
out of the way.

Back to comp.lang.haskell | Previous | NextPrevious in thread | Next in thread | Find similar | Unroll thread


Thread

arrays instead of lists Sean McIlroy <namenobodywants@gmail.com> - 2012-02-20 00:49 -0800
  Re: arrays instead of lists Paul Rubin <no.email@nospam.invalid> - 2012-02-20 01:17 -0800
  Re: arrays instead of lists Dirk Thierbach <dthierbach@usenet.arcornews.de> - 2012-02-20 11:27 +0100

csiph-web