comp.lang.ada
 help / color / mirror / Atom feed
From: Simon Wright <simon@pushface.org>
Subject: Re: Timing code blocks
Date: Fri, 05 Mar 2010 20:17:27 +0000
Date: 2010-03-05T20:17:27+00:00	[thread overview]
Message-ID: <m2mxymttew.fsf@pushface.org> (raw)
In-Reply-To: 4b90bf24$0$1979$4f793bc4@news.tdc.fi

I found the same results as Niklas, the compiler I'm running
is gcc version 4.5.0 20100221 (experimental) [trunk revision 156937]
(GCC) for x86_64-apple-darwin10.2.0.

I just added

   Junk_In    : Natural := 0;
   pragma Volatile (Junk_In);       -- <=====

after which -O[0123] all gave pretty similar results.

I didn't have to add 'delay 0.0;' to get the timings.

I did set

   Test_Dur : constant := 10_000;

and the timings on this Macbook Pro (2.4 GHz Core 2 Duo are of order
0.292 with -gnato, 0.254 without. The second loop is quicker with -O0.



      reply	other threads:[~2010-03-05 20:17 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-03-05  6:34 Timing code blocks deadlyhead
2010-03-05  7:55 ` Dmitry A. Kazakov
2010-03-05  8:16   ` deadlyhead
2010-03-05  8:49     ` Dmitry A. Kazakov
2010-03-05 12:41       ` Alex Mentis
2010-03-05 21:15         ` Jeffrey R. Carter
2010-03-05 23:35         ` Simon Wright
2010-03-06  9:50           ` Georg Bauhaus
2010-03-06 12:06             ` Simon Wright
2010-03-07  1:02               ` Georg Bauhaus
2010-03-08 12:16                 ` Alex Mentis
2010-03-06 12:12           ` Alex Mentis
2010-03-05  8:33   ` Ludovic Brenta
2010-03-05  9:04     ` Dmitry A. Kazakov
2010-03-05 15:27       ` Reporting bugs (was: Timing code blocks) Ludovic Brenta
2010-03-06  7:13         ` Stephen Leake
2010-03-05 23:35   ` Timing code blocks Simon Wright
2010-03-05  8:21 ` Niklas Holsti
2010-03-05 20:17   ` Simon Wright [this message]

replies disabled
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox