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: a07f3367d7,904e84d5dc020db1
X-Google-Attributes: gida07f3367d7,public,usenet
X-Google-NewGroupId: yes
X-Google-Language: ENGLISH,ASCII-7-bit
X-Received: by 10.66.76.201 with SMTP id m9mr59801paw.33.1361489656107;
        Thu, 21 Feb 2013 15:34:16 -0800 (PST)
Path: 
 ov8ni13556pbb.1!nntp.google.com!border1.nntp.dca.giganews.com!border4.nntp.dca.giganews.com!border2.nntp.dca.giganews.com!nntp.giganews.com!novia!news-hub.siol.net!news.mi.ras.ru!goblin-spool!goblin1!goblin.stu.neva.ru!ottix-news.ottix.net!newsswitch.lcs.mit.edu!nntp.TheWorld.com!.POSTED!not-for-mail
From: Robert A Duff <bobduff@shell01.TheWorld.com>
Newsgroups: comp.lang.ada
Subject: Re: Question about concurrent access to array
Date: Thu, 14 Feb 2013 16:53:57 -0500
Organization: The World Public Access UNIX, Brookline, MA
Message-ID: <wcchale5zpm.fsf@shell01.TheWorld.com>
References: <947b0951-54e4-4a4d-bd26-760eb66390c3@googlegroups.com>
 <bf027023-8b59-4d12-a077-d2e757a5a58d@googlegroups.com>
NNTP-Posting-Host: shell01.theworld.com
Mime-Version: 1.0
X-Trace: pcls6.std.com 1360878837 23423 192.74.137.71 (14 Feb 2013 21:53:57
 GMT)
X-Complaints-To: abuse@TheWorld.com
NNTP-Posting-Date: Thu, 14 Feb 2013 21:53:57 +0000 (UTC)
User-Agent: Gnus/5.1008 (Gnus v5.10.8) Emacs/21.3 (irix)
Cancel-Lock: sha1:pKOVWmtAPlZEp1TMZ8tVSn6yw/c=
Content-Type: text/plain; charset=us-ascii
Date: 2013-02-14T16:53:57-05:00
List-Id: <comp.lang.ada>

Adam Beneschan <adam@irvine.com> writes:

> I'd use an alignment aspect (or representation clause) to make sure
> that the array elements are aligned on some boundary, probably based
> on the largest integer that your machine can deal with.

That's not necessary for correctness.  It might be necessary to avoid
false sharing of cache lines, as Niklas Holsti mentioned.  Depends on
the machine.

>...Otherwise
> there's the risk that two consecutive array elements that "belong" to
> different tasks might share part of a "word" (or something), and the
> code to write into an element might involve reading the entire word,
> modifying part of the word, then writing it back.

That would be incorrect code generation, assuming there's no Pack
or Component_Size clause.

- Bob