caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: Eray Ozkural <exa@kablonet.com.tr>
To: Neel Krishnaswami <neelk@alum.mit.edu>,
	"caml Mailing List'" <caml-list@inria.fr>
Subject: Re: [Caml-list] Efficiency of 'a list
Date: Sun, 4 May 2003 15:56:06 +0300	[thread overview]
Message-ID: <200305041556.06626.exa@kablonet.com.tr> (raw)
In-Reply-To: <16052.61901.189814.21042@h00045a4799d6.ne.client2.attbi.com>

On Sunday 04 May 2003 13:56, Neel Krishnaswami wrote:
> > Don't trust microbenchmarks too far over what your knowledge of how
> > things should work tell you. Iterating over arrays is certainly
> > going to be much more cache-and TLB-friendly.
>
> This is not be as true you think. Ocaml's garbage collector is a
> compacting, copying GC, so it's very likely that lists will end up in
> in continuous blocks of memory. This will end up being nearly as
> cache-friendly as an array is.
>
> The big exception is with arrays of floats -- Ocaml unboxes arrays of
> floats, but doesn't unbox lists of them.

Now, we're getting some performance talk :)

To be precise, comparing an int list and int array we'll see that list 
occupies twice the same memory and therefore will generate more cache misses. 
But as the size of 'a in 'a list grows the difference will be negligible. 
However, if one modifies the elements of an array, in FORTRAN style, won't 
really be storing huge records inside elements of the array. He will likely 
split things up in parallel arrays where necessary, and if there are records 
they will be stored in a private heap and pointers will be stored in the 
array, etc.

Cheers,

-- 
Eray Ozkural (exa) <erayo@cs.bilkent.edu.tr>
Comp. Sci. Dept., Bilkent University, Ankara  KDE Project: http://www.kde.org
www: http://www.cs.bilkent.edu.tr/~erayo  Malfunction: http://mp3.com/ariza
GPG public key fingerprint: 360C 852F 88B0 A745 F31B  EA0F 7C07 AE16 874D 539C

-------------------
To unsubscribe, mail caml-list-request@inria.fr Archives: http://caml.inria.fr
Bug reports: http://caml.inria.fr/bin/caml-bugs FAQ: http://caml.inria.fr/FAQ/
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners


  reply	other threads:[~2003-05-04 12:56 UTC|newest]

Thread overview: 42+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2003-05-02 19:27 Eray Ozkural
2003-05-03  5:43 ` Mattias Waldau
2003-05-03  8:16   ` Ville-Pertti Keinonen
2003-05-03 14:12   ` Vitaly Lugovsky
2003-05-03 18:43     ` Mattias Waldau
2003-05-03 20:01       ` Eray Ozkural
2003-05-03 23:17       ` Eray Ozkural
2003-05-04  2:08       ` cashin
2003-05-04  4:08         ` alc
2003-05-04  5:32           ` Ed L Cashin
2003-05-04  6:46           ` [Caml-list] Two types of efficiency (Was Efficiency of 'a list) Mattias Waldau
2003-05-04  7:35             ` John Max Skaller
2003-05-04 11:52               ` Olivier Andrieu
2003-05-05 11:04                 ` John Max Skaller
2003-05-04 16:48               ` brogoff
2003-05-04  7:43             ` Ville-Pertti Keinonen
2003-05-04 12:50               ` Eray Ozkural
2003-05-04 12:48             ` Eray Ozkural
2003-05-05  7:31             ` Diego Olivier Fernandez Pons
2003-05-05 11:11               ` Mattias Waldau
2003-05-05 13:17                 ` John Max Skaller
2003-05-05 11:49               ` Eray Ozkural
2003-05-05 11:57               ` Yaron M. Minsky
2003-05-05 13:32                 ` John Max Skaller
2003-05-06  2:49                   ` Nicolas Cannasse
2003-05-06 12:30                     ` Diego Olivier Fernandez Pons
2003-05-07  2:05                       ` Nicolas Cannasse
2003-05-05 16:38                 ` Diego Olivier Fernandez Pons
2003-05-05 18:05                   ` Eray Ozkural
2003-05-06 13:28                     ` Diego Olivier Fernandez Pons
2003-05-13 11:35                   ` [Caml-list] Data Structure Libraries (was: Two types of efficiency) Oleg Trott
2003-05-04  7:55           ` [Caml-list] Efficiency of 'a list Ville-Pertti Keinonen
2003-05-04 10:56             ` Neel Krishnaswami
2003-05-04 12:56               ` Eray Ozkural [this message]
2003-05-04 13:35                 ` Falk Hueffner
2003-05-04 12:38           ` Eray Ozkural
2003-05-04  8:07         ` Ville-Pertti Keinonen
2003-05-04 15:54           ` Ed L Cashin
2003-05-05 23:52           ` Garry Hodgson
2003-05-03 20:03   ` Eray Ozkural
2003-05-03 21:13 ` Lauri Alanko
2003-05-03 22:03   ` Eray Ozkural

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200305041556.06626.exa@kablonet.com.tr \
    --to=exa@kablonet.com.tr \
    --cc=caml-list@inria.fr \
    --cc=erayo@cs.bilkent.edu.tr \
    --cc=neelk@alum.mit.edu \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).