[CPMD-list] Compilation of cpmd2cube

kohsj at ihpc.a-star.edu.sg kohsj at ihpc.a-star.edu.sg
Wed May 26 08:21:59 CEST 2004


Dear Dr. Axel,

Thanks, the file solved the problem.  Just one thing... towards the end of the compilation process, the following error message was printed:

xlf90 -o cpmd2cube.x -O -bmaxdata:32768000000 -bmaxstack:1500000000 kinds.o perl
ld: 0706-020 The number 32768000000 is too large for the -b maxdata option.     
        Use a value that is 4294967295 (hex FFFFFFFF) or less.                  
make: *** [cpmd2cube.x] Error 255                                               


I've followed the instructions and changed the -bmaxdata value to one order less (i.e. 3276800000).  Will that solve the problem, or are there other implications from this error message?

Thanks :)



----- Original Message -----
From: Axel Kohlmeyer <axel.kohlmeyer at theochem.ruhr-uni-bochum.de>
Date: Wednesday, May 26, 2004 1:38 pm
Subject: Re: [CPMD-list] Compilation of cpmd2cube

 
 On Tue, 25 May 2004 kohsj at ihpc.a-star.edu.sg wrote:
 
 AK> Dr. Axel,
 AK> 
 AK> I've used the attached "Makefile" to recompile.  The following 
 messageAK> appeared after "make" command was given:
 AK> 
 AK> 
 AK> make: *** No rule to make target `kinds.F', needed by `kinds.o'. 
 Stop.
 AK> 
 AK> 
 AK> Is there anything amiss?  Thanks!
 
 no, i messed up. the makefile was from a developer version 
 of cpmd2cube and the files have been renamed since.
 
 please try the attached re-redited version.
 
 sorry for the mess,
 	axel.
 
 AK> 
 AK> 
 AK> Adrian.
 AK> 
 AK> 
 AK> 
 AK> 
 AK> ----- Original Message -----
 AK> From: Axel Kohlmeyer <axel.kohlmeyer at theochem.ruhr-uni-bochum.de>
 AK> Date: Tuesday, May 25, 2004 5:01 pm
 AK> Subject: Re: [CPMD-list] Compilation of cpmd2cube
 AK> 
 AK>  On Tue, 25 May 2004 kohsj at ihpc.a-star.edu.sg wrote:
 AK>  
 AK>  AK> Dear all,
 AK>  AK> 
 AK>  AK> Thank you very much for all your helpful replies.  I'm 
 using IBM 
 AK>  p690AK> AIX 5L v.5.1 (Dr. Kohlmeyer is right).  I've followed your
 AK>  AK> instructions to delete and replace the necessary flags, and 
 theAK>  AK> "object files".  What was left with is this:
 AK>  
 AK>  noooooo. you should delete the .o files in your compile 
 directory not
 AK>  in the makesfile. hrmph, i always forget to word my replies 
 AK>  carefully...
 AK>  please try the attached makefile. this is for a newer version of
 AK>  cpmd2cube, but it should work nevertheless
 AK>  
 AK>  axel.
 AK>  
 AK>  AK> 
 AK>  AK> .SUFFIXES: .o .F90
 AK>  AK> 
 AK>  AK> ## IBM / AIX
 AK>  AK> FC	= xlf90 
 AK>  AK> FFLAGS	= -WF,"-D__AIX,-DFFT_ESSL" -qsuffix=cpp=F90 -O3 -
 AK>  qarch=pwr2 -qtune=pwr2
 AK>                                                                 
 ^^^^^^^AK>                                               this is the 
 main 
 AK>  culprit. 
 AK>     you have a power4 cpu in your machine, yet you tell the 
 compiler 
 AK>  to 
 AK>     compiler for power2 (exclusively)
 AK>  
 AK>  AK> LFLAGS	= -O2 -qtune=pwr2 -bmaxdata:1500000000 -
 bmaxstack:1500000000AK>                       ^^^^^^^^^^ 
 AK>                        this is wrong but not
 AK>  AK> LIBS	=  -L/rzu/aix/lib -lessl
 AK>  AK> #LIBS	= -L${HOME}/lib/fftw/lib -lfftw-double
 AK>  AK> 
 AK>  AK> ##############
 AK>  AK> 
 AK>  AK> EXE	= cpmd2cube
 AK>  AK> 
 AK>  AK> #############################
 AK>  AK> 
 AK>  AK> $(EXE):	
 AK>  AK> 	$(FC) -o $(EXE) $(LFLAGS) $(LIBS)
 AK>  AK> 
 AK>  AK> .F90.o:
 AK>  AK> 	$(FC) -c $(FFLAGS) $*.F90
 AK>  AK> 
 AK>  AK> clean: 
 AK>  AK> 	rm -f *.o *.mod
 AK>  AK> 
 AK>  AK> 
 AK>  AK> I've removed the ones for CRAY, HITACHI etc...  upon 
 AK>  recompiling, the following message came out:
 AK>  AK> 
 AK>  AK> 
 AK>  AK> xlf90  -o cpmd2cube -O2 -qtune=pwr2 -bmaxdata:1500000000 -
 AK>  bmaxstack:1500000000 l
 AK>  AK> ld: 0711-317 ERROR: Undefined symbol: .main                 
     
 AK>                
 AK>  AK> ld: 0711-345 Use the -bloadmap or -bnoquiet option to 
 obtain 
 AK>  more information.  
 AK>  AK> make: *** [cpmd2cube] Error 8                               
     
 AK>                
 AK>  AK> 
 AK>  AK> 
 AK>  AK> 
 AK>  AK> What was wrong?  Thanks again :).
 AK>  AK> 
 AK>  AK> 
 AK>  AK> Adrian.
 AK>  AK> 
 AK>  AK> 
 AK>  AK> 
 AK>  AK> 
 AK>  AK> 
 AK>  AK> ----- Original Message -----
 AK>  AK> From: Axel Kohlmeyer <axel.kohlmeyer at theochem.ruhr-uni-
 bochum.de>AK>  AK> Date: Tuesday, May 25, 2004 3:06 pm
 AK>  AK> Subject: Re: [CPMD-list] Compilation of cpmd2cube
 AK>  AK> 
 AK>  AK>  
 AK>  AK>  On Tue, 25 May 2004 kohsj at ihpc.a-star.edu.sg wrote:
 AK>  AK>  
 AK>  AK>  AK> Dear Sir,
 AK>  AK>  AK> 
 AK>  AK>  AK> Thanks a lot for clarifying on the compilation part.  
 I 
 AK>  managed to
 AK>  AK>  AK> compile the cpmd2cube by simply typing "make" within the
 AK>  AK>  AK> "cpmd2cube-v0.1.1" folder (with no changes made to the 
 AK>  cpmd2cubeAK>  AK> "Makefile").
 AK>  AK>  AK> 
 AK>  AK>  AK> Afterwhich, I've tried to convert the binary DENSITY 
 file 
 AK>  generatedAK>  AK> from the h2o exercise (this is in order to 
 display 
 AK>  the Kohn-Sham
 AK>  AK>  AK> Orbitals), using the following command:
 AK>  AK>  AK> 
 AK>  AK>  AK> > ./cpmd2cube DENSITY
 AK>  AK>  AK> 
 AK>  AK>  AK> The following error message came out:
 AK>  AK>  AK> 
 AK>  AK>  AK> Illegal instruction (core dumped)
 AK>  AK>  
 AK>  AK>  that usually means you have compiled for the wrong platform.
 AK>  AK>  my guess is, you are running on an AIX machine, right?
 AK>  AK>  
 AK>  AK>  if yes, please remove the -qarch=pwr2 flags,
 AK>  AK>  replace -lesslp2 with -lessl, delete all object files,
 AK>  AK>  and recompile.
 AK>  AK>  
 AK>  AK>  oh and olegs comment about havin to use -dens or -rho 
 AK>  AK>  applies as well.
 AK>  AK>  
 AK>  AK>  axel.
 AK>  AK>  
 AK>  AK>  AK> 
 AK>  AK>  AK> Why is this so?  The manual (S. 10.1.2) mentioned the 
 AK>  conversion of
 AK>  AK>  AK> WANNIER files only, am I missing or misinterpreting 
 AK>  anything?  
 AK>  AK>  ThanksAK> for your attention.
 AK>  AK>  AK> 
 AK>  AK>  AK> 
 AK>  AK>  AK> Adrian
 AK>  AK>  AK> 
 AK>  AK>  
 AK>  AK>  -- 
 AK>  AK>  
 AK>  AK>  
 AK>  AK>  
 AK>  
 =======================================================================AK>  Dr. Axel Kohlmeyer                        e-mail: axel.kohlmeyer at rub.de
 AK>  AK>  Lehrstuhl fuer Theoretische Chemie          Phone: ++49 
 AK>  (0)234/32-26673
 AK>  AK>  Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 
 AK>  (0)234/32-14045
 AK>  AK>  D-44780 Bochum  http://www.theochem.ruhr-uni-
 AK>  bochum.de/~axel.kohlmeyer/AK>  
 AK>  
 =======================================================================AK>  
 AK>  AK>  
 AK>  AK>  _______________________________________________
 AK>  AK>  CPMD-list mailing list
 AK>  AK>  CPMD-list at cpmd.org
 AK>  AK>  http://www.cpmd.org/mailman/listinfo/cpmd-list
 AK>  AK>  
 AK>  AK> 
 AK>  
 AK>  -- 
 AK>  
 AK>  
 AK>  
 =======================================================================AK>  Dr. Axel Kohlmeyer                        e-mail: axel.kohlmeyer at rub.de
 AK>  Lehrstuhl fuer Theoretische Chemie          Phone: ++49 
 (0)234/32-26673
 AK>  Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 
 (0)234/32-14045
 AK>  D-44780 Bochum  http://www.theochem.ruhr-uni-
 bochum.de/~axel.kohlmeyer/AK>  
 =======================================================================AK>  
 AK>  
 AK> 
 
 -- 
 
 
 =======================================================================
 Dr. Axel Kohlmeyer                        e-mail: axel.kohlmeyer at rub.de
 Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-26673
 Ruhr-Universitaet Bochum - NC 03/53         Fax:   ++49 (0)234/32-14045
 D-44780 Bochum  http://www.theochem.ruhr-uni-bochum.de/~axel.kohlmeyer/
 =======================================================================
 
 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: kohsj.vcf
Type: text/x-vcard
Size: 255 bytes
Desc: Card for <kohsj at ihpc.a-star.edu.sg>
Url : http://cpmd.org/pipermail/cpmd-list/attachments/20040526/b6c931b0/attachment.bin 


More information about the CPMD-list mailing list