Mailing-List: contact cygwin-help AT cygwin DOT com; run by ezmlm
List-Subscribe: <mailto:cygwin-subscribe AT cygwin DOT com>
List-Archive: <http://sourceware.org/ml/cygwin/>
List-Post: <mailto:cygwin AT cygwin DOT com>
List-Help: <mailto:cygwin-help AT cygwin DOT com>, <http://sourceware.org/ml/#faqs>
Sender: cygwin-owner AT cygwin DOT com
Mail-Followup-To: cygwin AT cygwin DOT com
Delivered-To: mailing list cygwin AT cygwin DOT com
To: cygwin AT cygwin DOT com
From: Mark Hadfield <m DOT hadfield AT niwa DOT co DOT nz>
Subject: Re: Memory for large arrays in cygwin/g77
Date: Tue, 11 Jan 2005 11:19:17 +1300
Lines: 38
Message-ID: <cruv1a$87f$1@sea.gmane.org>
References: <crurea$eg8$1 AT sea DOT gmane DOT org> <NUTMEG6Tnhgz2Uogux400000e95 AT NUTMEG DOT CAM DOT ARTIMI DOT COM>
Mime-Version: 1.0
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: 7bit
X-Complaints-To: usenet AT sea DOT gmane DOT org
X-Gmane-NNTP-Posting-Host: clam.niwa.cri.nz
User-Agent: Mozilla Thunderbird 1.0 (Windows/20041206)
In-Reply-To: <NUTMEG6Tnhgz2Uogux400000e95@NUTMEG.CAM.ARTIMI.COM>
X-IsSubscribed: yes

Dave Korn wrote:
>>From: cygwin-owner On Behalf Of Mark Hadfield
>>Sent: 10 January 2005 21:18
> 
>>Is there a cure that would allow a simple-minded, grey-haired Fortran 
>>programmer like me to rely on Cygwin g77 (or gfortran) for 
>>moderate-sized computational tasks?
> 
> LOL, I love the subtle sense of understatement you fortran guys use to
> describe handling arrays of hundreds of millions of elements :)

Though I am prone to occasionally employing subtlety, understatement and 
  irony for effect in forums like this, my reference to moderate-sized 
computational tasks in this connection was not such an occasion.

I run an ocean hydrodynamic model (ROMS) with domain size around 200 x 
200 x 30, sometimes larger. For the serious computations I use a Cray 
T3E, but for smaller stuff I use my Windows desktop machine or one of 
several Linux x86 machines with 1 or 2 CPUs, whatever I can get my hands 
on. For me a moderate-size simulation needs 256-512 MiB of memory. (If 
they're much larger than that, then they run too slow on the PC, so the 
limits associated with 32-bit addressing are not a major issue for me 
yet, but they will be in a few years' time.)

 > Anyway, the simple answer appears to be "Use g95 instead",
 > doesn't it?

Bascally, yes. I've been using g95 recently and have found it reliable 
and able to handle moderate-sized simulations, though execution is 
rather slow. However I did get a "MapViewOfFileEx" error recently with 
g95 on an old version of ROMS. That's not a show-stopper, as I seldom 
use that version any more, but it's a little worrying...


-- 
Mark Hadfield            "Ka puwaha te tai nei, Hoea tatou"
m DOT hadfield AT niwa DOT co DOT nz
National Institute for Water and Atmospheric Research (NIWA)


--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Problem reports:       http://cygwin.com/problems.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/