From chemistry-request@server.ccl.net  Sun Aug  8 22:54:26 1999
Received: from fg702-6.abct.polyu.edu.hk (fg702-6.abct.polyu.edu.hk [158.132.74.201])
	by server.ccl.net (8.8.7/8.8.7) with ESMTP id WAA13319
	for <CHEMISTRY@server.ccl.net>; Sun, 8 Aug 1999 22:54:25 -0400
Received: from fg702-6.abct.polyu.edu.hk (dannt.polyu.edu.hk [158.132.31.49])
	by fg702-6.abct.polyu.edu.hk (8.8.8/8.8.8) with ESMTP id KAA01582
	for <CHEMISTRY@www.ccl.net>; Mon, 9 Aug 1999 10:50:14 +0900 (HKST)
Message-ID: <37AE40D3.6D70D31B@fg702-6.abct.polyu.edu.hk>
Date: Mon, 09 Aug 1999 10:45:39 +0800
From: Daniel Mok <dkwmok@fg702-6.abct.polyu.edu.hk>
X-Mailer: Mozilla 4.61 [en] (Win95; I)
X-Accept-Language: en,zh-TW,zh-CN
MIME-Version: 1.0
To: CCL <CHEMISTRY@server.ccl.net>
Subject: Summary of Compilation in Linux using g77
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Dear all,

I have received several requests to ask me to summarize the response on
the question I posted last few weeks. I had a holiday, thus this is a
bit late... 8-)

Here is the original question:
> 
> I wonder has anyone tried to compile G94 or G98 under Linix using g77.
> I have tried to use f2c and successfully compile G94, but some
> executables do not function properly. While compiling G98 using f2c 
> failed. I would appreciate any suggestion. Thanks
> 

Several people suggested me to use the DEMO version of Portland F77
compiler with the blas-f2c.a available on the Gaussian Inc. www site.
Thank you for them.

If for some reason, you really want to compile G98 or G94 using g77,
though I have tried the solution they may help:

For G98 from Fred P. Arnold.
----------------------------
>If you look in mdutils.c there are statements for two commands, wrapped
>inside an #ifdef block, marked vfork and ?????.  If you use g77 as your
>base compiler, you have to make sure they get defined, since they >aren't in the default libraries.
>
>After that, -malign-doubles -O2, and start fooling with the various
>compiler flags until it passes the tests. 
>
>More importantly, be sure that you use EGCS 1.2 or later for your >version of g77.  It often takes more steps to converge, but will give >the same answer as that given by the Portland Group compiler.

For G94 from Darko Babic
-------------------------
>        I have compiled G94 under Linux with g77.  It was necessary to >modify the original source at few places, but I didn't notice any 
>improvement in comparison with f2c (in fact f2c version was slightly 
>faster).

>        If you want, I can send you patches for g77 compilation.  But I 
>wonder why your f2c version does not work properly.  You should look 
>into installation log file and see if there was any problems during 
>compilation and/or linking the links which are erratic.  I'd be glad to 
>help if you find any specific problem.

When I tried to use f2c to compile G94, f2c has created all the
executables for G94 successfully. However, some of the links terminated
improperly without giving any error message. As far as I remember, it is
either l511.exe or l512.exe failed. Thus, some of the SCF calculation
cannot finished but some of them can depend of the excutation route.
Any guy tried the same thing under RedHat 6.0 and have segmentation
fault reported. Anyone knows the reason?

Lastly, thank you for all your advices.

Regards,
Daniel
-- 
Office: (852)-2766 5629
Fax:    (852)-2364 9932
Department of Applied Biology and Chemical Technology
The Hong Kong Polytechnic University.
