AMBER Archive (2009)

Subject: Re: [AMBER] Test run on Solaris 10 X 86

From: Manikandan Chandrasekaran (mxchandrasek_at_ualr.edu)
Date: Tue Feb 03 2009 - 14:28:37 CST


I am sorry. I should have posted the error.

I am getting the following error, when I used ./configure_at solaris_cc

I happened to place a symbolic link for suncc, with the name solaris_cc

I downloaded Sun studio 12.

 "boundsutil.c", line 1121: warning: implicit function declaration: countmolatoms
 "boundsutil.c", line 1267: warning: implicit function declaration: NAB_ari
 ...
 ...
 cc: Warning: illegal option -fnostd
 cc -c -fnostd -DSYSV -DBINTRAJ   -o database.o database.c
 cc: Warning: illegal option -fnostd
 "database.c", line 435: warning: argument #4 is incompatible with prototype:
     .......
 "lex.dg_options.c", line 75: non-constant initializer: op "NAME"
 cc: acomp failed for embed.c
 *** Error code 2
 make: Fatal error: Command failed for target `embed.o'
 Current working directory /usr/local/amber10/src/nab
 *** Error code 1
 make: Fatal error: Command failed for target `install'
 
thanks

Mani

----- Original Message -----
From: Atro Tossavainen <atro.tossavainen+amber_at_helsinki.fi>
Date: Monday, February 2, 2009 5:07 pm
Subject: Re: [AMBER] Test run on Solaris 10 X 86
To: AMBER Mailing List <amber_at_ambermd.org>

> > I was not sure of how to use those compilers with amber.
>
> configure_amber sparc
>
> (although this is not strictly speaking appropriate as yours is
> not a
>  SPARC box but the compiler options for Sun Studio on SPARC
> and Sun
>  Studio on x86(_64) are remarkably similar, surprisingly
> enough, so
>  it should be a decent starting point)
>
> > The sun studio has suncc, sunf77, sunf90, sunf95 and so forth.
>
> It ought to have just cc, f77, f90, f95 and so forth without any
> prefixesat all.  At least here it does.
>
> NB I've only got Solaris on SPARC, haven't bothered with the x86
> kind so far.
>
> > There are many compilers in it. With these many compilers,
> which one
> > should be used with the configure option ?
>
> Regarding AMBER itself:
>
> You can't compile FORTRAN code with cc, and you can't compile AMBER
> without a FORTRAN-90 capable compiler.  On Studio 11 at
> least, f95
> is just a link to f90.  That leaves you with exactly one compiler
> that will compile the main body of AMBER.
>
> Regarding AmberTools:
>
> The configuration option solaris_cc picks the right C and FORTRAN
> compilers and options for you on both Solaris x86 and Solaris SPARC.
>
> In general:
>
> AMBER's configure scripts have been written to take into account
> whateverany particular vendor calls their compilers.  The
> default in configure_amber
> is, for example, to use whatever "f90" leads to on any given
> system UNLESS
> explicitly configured otherwise.
>
> > I can comprehend on the fact that ambertools has an option for
> solariscc. So perhaps I can place a symbolic link for suncc with
> solariscc. ( I tried but it didn't work )
>
> Maybe you can post the error messages you get rather than just vaguely
> say "it didn't work" if you want help.
>
> > But for amber10, when I looked in for help, I am not able to
> see which compiler is actually refering to any of the compilers
> in sunstudio. There are options like ifort, pgf95...
>
> configure_amber sparc
>
> I can appreciate the fact that everybody must be a newbie at least
> once, but being able to post the explicit error messages you receive
> rather than "it doesn't work!!!" is not related to being a newbie.
>
> --
> Atro Tossavainen
> (Mr.)               / The Institute of Biotechnology at
> Systems Analyst, Techno-Amish &     / the
> University of Helsinki, Finland,
> +358-9-19158939  UNIX Dinosaur     /
> employs me, but my opinions are my own.
> < URL : http : / / www . helsinki . fi / %7E atossava / > NO
> FILE ATTACHMENTS
>
> _______________________________________________
> AMBER mailing list
> AMBER_at_ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber
_______________________________________________
AMBER mailing list
AMBER_at_ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber