Re: [AMBER] Error Compiling NAB o AMBER Tools 1.2

From: Giorgos Lamprinidis <lamprinidis.pharm.uoa.gr>
Date: Thu, 15 Jul 2010 17:34:59 +0300

Dear friends,

I thank you for your suggestions. I found out that....

There is some previous post on Jun 10 2010 from colvin who had the same
problem with a mac computer. However he didn;t found a solution. He just
compiled the 32bit version, since he realised that ther was some problem
with the new gcc 64bit compilers.
So i made some tests on other 64bit machines i have.
1) Quad Core Intel with gcc 4.4.1 20090725 9Red Hat 4.4.1-2), fedora core 11
(2.6.30.10-105.2.23.fc11.x86_64)
2) Dual Core Intel with gcc 4.3.0 20080428 (Red Hat 4.3.0-8) , fedora core 9
(2.6.27.25-78.2.56.fc9.x86_64)
On both machines AmberTools compiled without problems, (except tha addition
of #include <cstdio> on ./src/gleap/mortsrc/common/fortran.cpp file) and
passed all tests.

On the machine i have the error on nab module, the gcc version is 4.4.4
20100630 (Red Hat 4.4.4-10) on Fedora Core 13 2.6.33.6-147.fc13.x86_64
I manually converted
printf( "\"%s\" is not a valid answer.\n", ParallelAnswer );
to
printf("\"%s\" is not a valid answer.\n", ParallelAnswer );
(i erased the space between ( and " )
gcc compiled it without errors, but when i run the tests....nab failed
because the executable "nab" had errors. It means, it wasn't compiled
correctly.

Finally I downloaded from intel the icc compiler and i compiled ambertools
using icc..... it was like a miracle....nab was compiled without a problem
and passed all the tests. So.....can anyone tell me what is wrong with the
latest 64bit compilers of gcc? Have you seen something similar on the past?

Thanks again for your time.


George

----- Original Message -----
From: "case" <case.biomaps.rutgers.edu>
To: "AMBER Mailing List" <amber.ambermd.org>
Sent: Wednesday, July 14, 2010 11:26 PM
Subject: Re: [AMBER] Error Compiling NAB o AMBER Tools 1.2


> On Wed, Jul 14, 2010, Giorgos Lamprinidis wrote:
>>
>> I am trying to compile AmberTools 1.2 to a new Linux machine having 64bit
>> Intel Core2 vPro proc.
>> 1) I patch the source code.
>> 2) I run the configuration script using -gcc compiler.
>
>>
>> When i search the na_anal.c file i see at line 2417 the following:
>> 2416 else{
>> 2417 printf( "\"%s" is not a valid answer.\n", ParallelAnswer );
>> 2418 NotValidAnswer = 1;
>> 2419 }
>
> I don't see this: my equivalent line is
>
> printf( "\"%s\" is not a valid answer.\n", ParallelAnswer );
>
> Please report the results of typing "gcc --version" and "uname -a". Also
> check that na_anal.nab has the following line (line 2188):
>
> printf( "\"%s\" is not a valid answer.\n", ParallelAnswer );
>
> The nab2c program is just passing this line straight through to the C
> compiler (or should be).
>
> [I vaguely, vaguely remember something like this popping up before, but
> don't see anything in the archives about it...maybe someone else has a
> better
> memory....]
>
> ....dac
>
>
> _______________________________________________
> AMBER mailing list
> AMBER.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
>


_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Jul 15 2010 - 08:00:04 PDT
Custom Search