From mboxrd@z Thu Jan  1 00:00:00 1970
X-Spam-Checker-Version: SpamAssassin 3.4.4 (2020-01-24) on polar.synack.me
X-Spam-Level: 
X-Spam-Status: No, score=-1.9 required=5.0 tests=BAYES_00 autolearn=ham
	autolearn_force=no version=3.4.4
X-Google-Thread: 103376,4cf1fd41f64f8f02
X-Google-Attributes: gid103376,public
X-Google-Language: ENGLISH,ASCII-7-bit
Path: 
 g2news2.google.com!news3.google.com!news.glorb.com!atl-c08.usenetserver.com!news.usenetserver.com!pc03.usenetserver.com!news.flashnewsgroups.com-b7.4zTQh5tI3A!not-for-mail
Newsgroups: comp.lang.ada
Subject: Re: task-safe hash table?
References: <uverhhygc.fsf@acm.org>
 <FradnWbzpPdOix7ZnZ2dnUVZ_qmdnZ2d@comcast.com>
From: Stephen Leake <stephen_leake@acm.org>
Date: Wed, 07 Jun 2006 21:10:26 -0400
Message-ID: <uodx4jvul.fsf@acm.org>
User-Agent: Gnus/5.1006 (Gnus v5.10.6) Emacs/21.3 (windows-nt)
Cancel-Lock: sha1:QcEIjb5IsCkN1+rR01hSQtkRj54=
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Complaints-To: abuse@flashnewsgroups.com
Organization: FlashNewsgroups.com
X-Trace: acf8c44877910e73ae4a425912
Xref: g2news2.google.com comp.lang.ada:4706
Date: 2006-06-07T21:10:26-04:00
List-Id: <comp.lang.ada>

tmoran@acm.org writes:

> with ada.calendar,
>      ada.text_io;
> procedure timeprot is
>
> <snip>

This looks like a reasonable test. Here's my results with GNAT 5.04 on
Windows XP (I only kept the second run; the first loads cache):

gnatmake -O0
./timeprot.exe 
a 0.025702000 mics
b 0.026539000 mics
c 0.029333000 mics
prot.a 4.420673000 mics
prot.b 0.504533000 mics
prot.c 0.423238000 mics


gnatmake -O2
./timeprot.exe 
a 0.023188000 mics
b 0.022629000 mics
c 0.023187000 mics
prot.a 7.770820000 mics
prot.b 0.576610000 mics
prot.c 0.529676000 mics

Interesting that it is slower with -02!

AdaCore recommends against using -O3 because it includes "potentially
unsafe and disruptive optimizations". But here it is anyway:

gnatmake -O3
./timeprot.exe 
a 0.014806000 mics
b 0.014527000 mics
c 0.004470000 mics
prot.a 7.431950000 mics
prot.b 0.619911000 mics
prot.c 0.529676000 mics

There is some variation from run to run. But this is pretty convincing
that a protected function call is 25 times slower than a plain
function call. I'll see what AdaCore has to say. And I'll try it on
Lynx, which is more real-time than Windows or Linux :).

-- 
-- Stephe