From mboxrd@z Thu Jan  1 00:00:00 1970
Received: by pauillac.inria.fr; Wed, 16 Feb 94 20:08:00 +0100
Received: from margaux.inria.fr by pauillac.inria.fr; Wed, 16 Feb 94 14:32:34 +0100
Received: from concorde.inria.fr by margaux.inria.fr, Wed, 16 Feb 94 14:32:31 +0100
Received: from stroma.dcs.ed.ac.uk by concorde.inria.fr; Wed, 16 Feb 1994 14:32:26 +0100
Received: from whalsay.dcs.ed.ac.uk by dcs.ed.ac.uk id aa28655;
          16 Feb 94 13:26 GMT
Date: Wed, 16 Feb 94 13:26:45 GMT
Message-Id: <10728.9402161326@whalsay.dcs.ed.ac.uk>
From: Christophe Raffalli <cr@dcs.ed.ac.uk>
Sender: weis@pauillac.inria.fr
To: caml-list@margaux.inria.fr
Subject: Times in camllight


I am trying to use the times function from the unix library. I remark quite
big variation (for utime, stime and the sum) when I measure many times the same
computation (starting myself a gc with "gc();gc();" before each computation).

Here is an example of this variation :

unification : times = 1.01666666667+0.133333333333=1.15
unification : times = 0.666666666667+0.15=0.816666666667
unification : times = 1.13333333333+0.166666666667=1.3
unification : times = 1.01666666667+0.233333333333=1.25
unification : times = 0.65+0.116666666667=0.766666666667

min = 0.766666666667
max = 1.3

!!!

Why ???

Christophe

--
----------------------------------------------------------
Christophe Raffalli
e-mail: cr@dcs.ed.ac.uk

L.F.C.S. (Laboratory for Fundation of Computer Science)
Department of Computer Science, University of Edinburgh
King's Building, Mayfield Road, Edinburgh EH9 3JZ
----------------------------------------------------------