caml-list - the Caml user's mailing list
 help / color / mirror / Atom feed
From: "Jon Harrop" <jon@ffconsultancy.com>
To: "'Tom Ridge'" <tom.j.ridge+list@googlemail.com>,
	"'caml-list'" <caml-list@inria.fr>
Subject: RE: [Caml-list] Question about garbage collection and impact on performance
Date: Wed, 4 Dec 2013 18:09:31 -0000	[thread overview]
Message-ID: <00cd01cef11b$fbe6c8e0$f3b45aa0$@ffconsultancy.com> (raw)
In-Reply-To: <CABooLwP7eSJ5Zc=3uSwCqU4a8Yy6a8bSm_74fQEhN=ceEzpdVw@mail.gmail.com>

> is it possible for OCaml's garbage collection to alter the time complexity
of my program?

Yes.

> If the answer is "yes", then are there any expectations I might have about
how bad this alteration might be?

I observed List.map taking O(n^2) time for large "n" because it leaks stack
space and the GC traverses the O(n)-deep stack O(n) times during its
execution.

> For example, (without thinking about it too hard) it seems unreasonable
for GC to turn a polytime program into an exponential time program, but is
this actually true for OCaml's garbage collector?

I think you can just get an extra O(n). I'd also look out for very large
arrays on the heap which, IIRC, cause O(n) pause.

I haven't used OCaml in anger for a few years so take this with a grain of
salt... :-)

Cheers,
Jon.

-----Original Message-----
From: tom.j.ridge@googlemail.com [mailto:tom.j.ridge@googlemail.com] On
Behalf Of Tom Ridge
Sent: 04 December 2013 12:21
To: caml-list
Subject: [Caml-list] Question about garbage collection and impact on
performance

Dear caml-list,

I have an OCaml program which I expect to run in time O((n^3) *
(ln(n))) say. My expectations are based (unrealistically) on ignoring
garbage collection completely. As inputs get large, the program performs
worse than I expect.

My question is: is it possible for OCaml's garbage collection to alter the
time complexity of my program?

If the answer is "yes", then are there any expectations I might have about
how bad this alteration might be? For example, (without thinking about it
too hard) it seems unreasonable for GC to turn a polytime program into an
exponential time program, but is this actually true for OCaml's garbage
collector?

I have read some material on OCaml's GC, but I did not form any intuitions
yet about the answers to these questions.

Thanks in advance

Tom

--
Caml-list mailing list.  Subscription management and archives:
https://sympa.inria.fr/sympa/arc/caml-list
Beginner's list: http://groups.yahoo.com/group/ocaml_beginners
Bug reports: http://caml.inria.fr/bin/caml-bugs


  parent reply	other threads:[~2013-12-04 18:09 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2013-12-04 12:20 Tom Ridge
2013-12-04 12:43 ` Malcolm Matalka
2013-12-04 13:48 ` Gabriel Scherer
2013-12-19 22:47   ` Richard W.M. Jones
2013-12-22  2:25     ` Jon Harrop
2013-12-22  3:04       ` David Sheets
2013-12-22 12:43         ` Jeff Schultz
2013-12-22 10:41       ` Richard W.M. Jones
2013-12-04 18:09 ` Jon Harrop [this message]
2013-12-05  1:09 ` Jacques Garrigue
2013-12-05 10:08   ` Tom Ridge
2013-12-05 10:37     ` Malcolm Matalka
2013-12-05 10:48     ` Michael Hicks
2013-12-05 11:21       ` Tom Ridge
2013-12-05 21:09       ` Jon Harrop

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='00cd01cef11b$fbe6c8e0$f3b45aa0$@ffconsultancy.com' \
    --to=jon@ffconsultancy.com \
    --cc=caml-list@inria.fr \
    --cc=tom.j.ridge+list@googlemail.com \
    /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).