AMBER Archive (2009)
Subject: Re: [AMBER] problem compiling nmode.c
From: Shan-ho Tsai (tsai_at_hal.physast.uga.edu)
Date: Thu Feb 19 2009 - 08:49:39 CST
Thank you very much, Dr Case.
After implementing the changes you described, I was able
to compile AmberTools and all tests were fine. I had to
compile lmodC.c with optimization turned off, i.e -O0,
because with -O the minimum energy for the simple lmod
optimization test did not fall within -115 to -125, but
with -O0 the value turned out to be -123.
I really appreciate your kind help.
Thank you again,
On Tue, 17 Feb 2009, David A. Case wrote:
> On Tue, Feb 17, 2009, Shan-ho Tsai wrote:
>> I'm trying to build AmberTools 1.2 on an AIX 5.2 box,
>> using gcc 3.3.2 (gcc 4.1.2 gave the same errors), and
>> I am getting an error compiling nmode.c:
>> gcc -c -DBINTRAJ -o nmode.o nmode.c
>> nmode.c: In function `lnorm':
>> nmode.c:1172: error: `__I' undeclared (first use in this function)
>> nmode.c:1172: error: (Each undeclared identifier is reported only once
>> nmode.c:1172: error: for each function it appears in.)
>> make: 1254-004 The error code from the last command is 1.
>> The system does have /usr/include/complex.h, but it
>> must be non-standard.
>> So I added -DUSE_AMBER_C9XCOMPLEX to the compilation
>> flag to use the local complex.h and then I got many other
>> gcc -c -DBINTRAJ -DUSE_AMBER_C9XCOMPLEX -o nmode.o nmode.c
>> In file included from nmode.c:6:
>> complex.h:101: error: parse error before '(' token
>> complex.h:101: error: conflicting types for `atan2'
> Certainly is weird. A workaround: comment out the lnorm() function, the
> one call to it, and the include of complex.h. This should get the basic
> nmode functionality working. We'll try to get a version of nmode done
> soon that does not use complex arithmetic.
> AMBER mailing list
AMBER mailing list