Dear forum members
A colleague of mine is attempting to install Amber 11 onto an elderly Cray system and has run into the previously-documented error associated with xray_interface.f. He has tried the suggested fixes but to no avail. He is of the opinion that if he could obtain a copy of the file "xray_interface.f" from a working Amber 11 setup verified to work with PGI7.1-1 then he could carry on.
Is there anyone who could send me a copy of the file "xray_interface.f" verified to work with PGI7.1-1?
His message is appended to this email.
Thanks in advance.
Dan Rathbone
________________________________
Dr Dan Rathbone
School of Life and Health Sciences
Aston University
Aston Triangle
Birmingham
UK
________________________________
> Hi Dan,
>
> after a few attempts I have managed to build parallel version of
> AmberTools and started building Amber 11, until I came across this PGI
> (Fortlan90 compiler bug) this is how I found it.
> Currently Loaded Modulefiles:
> 1) pgi/7.1-1 2) mpich/mpich-pgi711
>
> This is the error.
>
> mpif90 -c -O1 -Mfree -o xray_globals.o _xray_globals.f cpp
> -traditional -P -DMPI xray_utils.f > _xray_utils.f mpif90 -c -O1
> -Mfree -o xray_utils.o _xray_utils.f
> PGF90-S-0000-Internal compiler error. size_ast_of: unexpected ast type
> 4 (_xray_utils.f: 567)
> 0 inform, 0 warnings, 1 severes, 0 fatal for write_header
> make[1]: *** [xray_utils.o] Error 2
> make[1]: Leaving directory `/data/apps/pgi64/amber_new/amber11/src/sander'
> make: *** [parallel] Error 2
>
> Building serial version produces the same error.
>
> I have found that the error is known to amber developers since January
> of 2009 it's in sander directory in file xray_interface.f here is the
> snippet: where is starts.
>
> ! GMS: Fix for pgf90 compiler
> character(len=4) :: this_residue_chainid
> ! begin
>
> and further down.
>
>
> ! GMS: Fix for pgf90 compiler
> this_residue_chainid = residue_chainid(ires)
>
> write(unit,'(A6,I5,1X,A4,A1,A3,1X,A1,I4,A1,3X,3F8.3,2F6.2,6X,2A4)')&
> merge('ATOM
> ','HETATM',any(resName==standard_pdb_residues)), &
> iatom,name,atom_altloc(iatom)(1:1), &
> resName,residue_chainid(ires)(1:1), &
> residue_number(ires),residue_icode(ires)(1:1), &
> coordinate(1:3,iatom), &
> atom_occupancy(iatom), &
> atom_bfactor(iatom), &
> ! GMS: Fix for pgf90 compiler
> !merge(residue_chainid(ires),' ',pdb_use_segid), &
> merge(this_residue_chainid,' ',pdb_use_segid), &
> atom_element(iatom)
> Looking at forum pages
> It appears that the problem was solved through this changes above, but
> I am getting the same error as if fix was not there My Fortran is non
> existent. I was going to suggest that if you purchase the licence and
> support , then perhaps you might ask for xray_interface.f file from
> amber guys (it's in a sander source) which would have all the correct
> fixes for PGI7.1-1. Then I believe I can build it.
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Thu Nov 18 2010 - 07:00:04 PST
