[CPMD-list] A final(?) linking problem..
Axel Kohlmeyer
akohlmey at cmm.chem.upenn.edu
Mon Jul 16 19:39:40 CEST 2007
On Mon, 16 Jul 2007, Dave Sherman wrote:
DS> Dear Juerg,
DS>
DS> Thank you for your quick response. That got me past fileopen. I
DS> also downloaded the libatlas_x86_64.a library
DS> to my source directory and was able to compile everything except that
DS> at the very end, it could not create the cpmd.x
DS> because the linking failed with
DS> .
DS> .
DS> shop_adds.o sysdepend.o -Bstatic -L. -latlas
DS> /usr/bin/ld: cannot find -libt
DS> make: *** [cpmd.x] Error 2
this looks like an unwanted linebreak somewhere. also,
please remove -Bstatic. on newer linux machines, fully
static linking can have very bad side effects.
DS> I'm confused about what libt is. I think it refers to libt.so.
no. with static linkinging the linker would look for libt.a,
but that aside, cpmd does not use/need a libt.a or libt.so
cheers,
axel.
DS> There is nothing
DS> about it in the Makefile. Any suggestions?
DS>
DS> Thanks for your help.
DS>
DS> Dave Sherman
DS>
DS> On 16 Jul 2007, at 16:28, Juerg Hutter wrote:
DS>
DS> > Hi
DS> >
DS> > your mpif77 probably really calls the pgf77 compiler and
DS> > not pgf90 as is assumed in this CONFIGURE file.
DS> > If you replace mpif77 by mpif90 in the Makefile it should work.
DS> >
DS> > regards
DS> >
DS> > Juerg Hutter
DS> >
DS> > ----------------------------------------------------------
DS> > Juerg Hutter Phone : ++41 44 635 4491
DS> > Physical Chemistry Institute FAX : ++41 44 635 6838
DS> > University of Zurich E-mail: hutter at pci.uzh.ch
DS> > Winterthurerstrasse 190
DS> > CH-8057 Zurich, Switzerland
DS> > ----------------------------------------------------------
DS> >
DS> >
DS> > On Mon, 16 Jul 2007, Dave Sherman wrote:
DS> >
DS> >> I'm trying to compile CPMD-3.11.1 using the PGI-AMD64-MPI
DS> >> configuration but I'm having
DS> >> trouble with the fileopen.F/fileopen.f subroutine:
DS> >> .
DS> >> .
DS> >> .
DS> >> rm -f fileopen.f
DS> >> /lib/cpp -P -C -traditional -D__Linux -D__PGI -DFFT_DEFAULT -
DS> >> DPOINTER8 -D__pgf90 -DPARALLEL -DMYRINET ./fileopen.F ./fileopen.f
DS> >> mpif77 -c -fastsse -tp k8-64 -r8 -pc=64 -Msignextend ./
DS> >> fileopen.f -o ./fileopen.o
DS> >> PGFTN-S-0038-Symbol, position, has not been explicitly declared (./
DS> >> fileopen.f: 110)
DS> >> PGFTN-S-0034-Syntax error at or near = (./fileopen.f: 110)
DS> >> 0 inform, 0 warnings, 2 severes, 0 fatal for fileopen
DS> >> make: *** [fileopen.o] Error 2
DS> >>
DS> >> In the source code fileopen.f,
DS> >> .
DS> >> ,
DS> >> ,
DS> >> IF(TFODBG) WRITE(*,*) 'FILEOPEN: OPEN FILE FOR APPENDING'
DS> >> OPEN(UNIT=IUNIT,FILE=FULLPATH(IA2:IE2),STATUS='OLD',
DS> >> & POSITION='APPEND',
DS> >> & FORM=FO_FORM,ERR=999)
DS> >> IF(TFOVRB) WRITE(*,'(3A)') ' FILE ',FILEN(IA1:IE1),
DS> >> & ' EXISTS, NEW DATA WILL BE APPENDED'
DS> >> ELSE
DS> >> .
DS> >> the offending keyword is POSITION which, apparently, the PGI
DS> >> compiler does not recognize. (I don't recognize it either..).
DS> >>
DS> >> Any suggestions on how to get around this?
DS> >>
DS> >> Thanks for your help.
DS> >>
DS> >> ______________
DS> >> David M. Sherman
DS> >> Professor of Geochemistry
DS> >> Department of Earth Sciences
DS> >> University of Bristol
DS> >> Bristol BS8 1RJ UNITED KINGDOM
DS> >> Phone: 44-(0)117-954-5446
DS> >> http://mineral.gly.bris.ac.uk
DS> >>
DS>
DS> David M. Sherman
DS> Professor of Geochemistry
DS> Department of Earth Sciences
DS> University of Bristol
DS> Bristol BS8 1RJ UNITED KINGDOM
DS> Phone: 44-(0)117-954-5446
DS> http://mineral.gly.bris.ac.uk
DS>
DS>
--
=======================================================================
Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu
Center for Molecular Modeling -- University of Pennsylvania
Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323
tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425
=======================================================================
If you make something idiot-proof, the universe creates a better idiot.
More information about the CPMD-list
mailing list