Re: [GC] Evaluate memory use

From: Matías Giovannini (matias@k-bell.com)
Date: Wed Nov 24 1999 - 19:39:22 MET


Date: Wed, 24 Nov 1999 15:39:22 -0300
From: =?iso-8859-1?Q?Mat=EDas?= Giovannini <matias@k-bell.com>
To: caml-list@inria.fr
Subject: Re: [GC] Evaluate memory use

Damien Doligez wrote:
>
> >From: David.Mentre@irisa.fr (David =?iso-8859-1?q?Mentré?=)
>
> >If I've understood the 2.02 doc, the compaction mechanism is disabled by
> >default. Right? So the below method is safe. Right?
>
> Yes.
>
> >Oh no. I've managed to use it. :) That's only because I must use the
> >Unix module solely on this purpose. It was just to avoid such use.
>
> Maybe some day, we'll have date and time functions in the standard
> library.
>
> -- Damien

Hm, I can see a new debate coming about the vagaries of calendrical
calculations around the world...

ObOCaml, it would be *extremely* useful if it only were

1- A system-independent time service, together with:
2- A system-independent time-base value (say, microseconds), and
3- A system-independent date-zero value (say, the datetime of release of
CamlLight 0.7).

What I'm thinking about is a sub-second TOD clock that could serve both
as a timer and as a clock. If the OS doesn't provide it, it's relatively
easy to sinthesize such a clock with a wall clock and a sub-second
timer. The critical point in doing it inside the language and not in the
library would be thread synchronization, but maybe I'm mistaken.

Best regards,
Matías

PS.: I just realized that the hypotetical Clock.t type would have to be
64 bits wide at least, that makes it unwieldly to manipulate inside OCaml.

-- 
I was seized by the hallucination. I don't remember much, except for
being caught in an infinite recursion: I was myself feverishly
writing how I was seized by the hallucination.  I don't remember
much, except for being caught in an infinite recursion.



This archive was generated by hypermail 2b29 : Sun Jan 02 2000 - 11:58:28 MET