From jkr@jkr.cc.rl.ac.uk Thu May 21 18:33:47 1998 Received: from nameserv.rl.ac.uk ([130.246.132.17]) by dkuug.dk (8.6.12/8.6.12) with ESMTP id SAA11164 for ; Thu, 21 May 1998 18:33:46 +0200 Received: from jkr.cc.rl.ac.uk (jkr.cc.rl.ac.uk [130.246.8.20]) by nameserv.rl.ac.uk (8.8.8/8.8.8) with ESMTP id RAA12729; Thu, 21 May 1998 17:28:30 +0100 Received: (from jkr@localhost) by jkr.cc.rl.ac.uk (8.8.8/8.8.8) id RAA18500; Thu, 21 May 1998 17:29:10 +0100 (BST) Date: Thu, 21 May 1998 17:29:10 +0100 (BST) From: John Reid Message-Id: <199805211629.RAA18500@jkr.cc.rl.ac.uk> To: SC22WG5@dkuug.dk, wg25@adm.csc.ncsu.edu, comp-fortran-90@mailbase.ac.uk Subject: Co-array Fortran Cc: rwn@cray.com X-Sun-Charset: US-ASCII Bob Numrich and I have been working on a paper on Co-array Fortran, formerly known as F--, over the last few months and have reached the point where we would appreciate comments from interested readers. A very brief summary of Co-array Fortran is to say that the same program runs asynchronously on all processes and arrays may be declared with up to 7 extra subscripts in square brackets, which are used to address data across processes. There are procedures to identify the index of the executing process and to synchronize processes. I think it provides a simple, elegant, and powerful way to express parallelism. I have put our draft paper in my ftp server: jkr.cc.rl.ac.uk in the file pub/co-array.ps.gz It is designed for A4 paper, so you may need to tell your printer. A nice informal article on F-- appeared in SIAM News: Numrich, R. W. and Steidel, J. L. (1997a). F--: A simple parallel extension to Fortran 90. SIAM News, <30>, 7, 1-8. This and a couple of other articles are available from http://www.cray.com/products/applications/fmm/ I look forward to hearing your comments. John Reid.