From owner-sc22wg5@open-std.org Fri Jun 17 23:03:01 2011 Return-Path: X-Original-To: sc22wg5-dom8 Delivered-To: sc22wg5-dom8@www2.open-std.org Received: by www2.open-std.org (Postfix, from userid 521) id 036EBC178E6; Fri, 17 Jun 2011 23:03:01 +0200 (CET DST) X-Original-To: sc22wg5@open-std.org Delivered-To: sc22wg5@open-std.org Received: from exprod6og101.obsmtp.com (exprod6og101.obsmtp.com [64.18.1.181]) by www2.open-std.org (Postfix) with ESMTP id DB11BC178DF for ; Fri, 17 Jun 2011 23:02:59 +0200 (CET DST) Received: from cfexcas02.americas.cray.com ([136.162.34.11]) (using TLSv1) by exprod6ob101.postini.com ([64.18.5.12]) with SMTP ID DSNKTfvBAveKBWKMQvXYUlW5vMNvHwddR9hk@postini.com; Fri, 17 Jun 2011 14:03:00 PDT Received: from fortran.us.cray.com (172.31.19.200) by cfexcas02.americas.cray.com (172.30.74.226) with Microsoft SMTP Server (TLS) id 8.3.137.0; Fri, 17 Jun 2011 16:02:57 -0500 Message-ID: <4DFBC103.9040703@cray.com> Date: Fri, 17 Jun 2011 16:02:59 -0500 From: Bill Long Reply-To: Organization: Cray Inc. User-Agent: Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.6; en-US; rv:1.9.2.17) Gecko/20110414 Thunderbird/3.1.10 MIME-Version: 1.0 To: sc22wg5 Subject: Re: (j3.2006) (SC22WG5.4481) [ukfortran] Added atomic References: <20110617160827.0A646C178E6@www2.open-std.org> <20110617205441.D8A9DC178DF@www2.open-std.org> In-Reply-To: <20110617205441.D8A9DC178DF@www2.open-std.org> Content-Type: text/plain; charset="ISO-8859-1"; format=flowed Content-Transfer-Encoding: 7bit Sender: owner-sc22wg5@open-std.org Precedence: bulk On 6/17/11 3:54 PM, N.M. Maclaren wrote: > On Jun 17 2011, Bill Long wrote: >> >> The list of comments on the coarray TR from Rice suggested adding a SWAP >> atomic subroutine. The list in 11-200 includes all of the "common" >> atomic operations except one: atomic andxor. I'm inclined to add that >> one, which would make SWAP unnecessary. Or we could add both. For those >> who are not big amo users... >> >> ... >> >> If this is added, is there a need for a separate atomic_swap? > > Swapping works reliably and portably when one integer may be negative; > XOR doesn't. Perhaps even more persuasively, the SWAP routine would be well defined for LOGICAL arguments (which we allow for _DEFINE and _REF), whereas the andxor version description is problematic for that case. So, I'm convinced that both versions are useful. Cheers, Bill > > Regards, > Nick. > > _______________________________________________ > J3 mailing list > J3@j3-fortran.org > http://j3-fortran.org/mailman/listinfo/j3 -- Bill Long longb@cray.com Fortran Technical Support & voice: 651-605-9024 Bioinformatics Software Development fax: 651-605-9142 Cray Inc./Cray Plaza, Suite 210/380 Jackson St./St. Paul, MN 55101