[Pw_forum] g95 compilation problem

[Silviu Zilberman]

Dear Paolo,

I do have a working version of Q-Espresso on cygwin/g95, attached is the 
make.sys file. Unfortunately I am not familiar enough with the configure 
utility to modify the configure.ac file myself.

The two major changes with respect to the original make.sys are:
1. The "-o" in the preprocessing step is not working properly and I had 
to use the redirection symbol instead.
2. There was a case problem. The original preprocessing is generating 
f90->F90. In cygwin it overwrites the original f90 file and all sort of 
strange things then happen. It seems to be a cygwin problem, ignoring 
the case of the extension. Therefore I changed the extension of the 
preprocessed file to F95 and added a rule for F95->object file.

I get executables that seem to run properly but I have one problem:
For some reason the path of (say) the pseudopotentials (PP) library is 
unrecognized by the code and it complains that it cannot fine the PP. 
The only way I found around it is to copy the PP to the local directory 
(symbolic link is not enough). The same problem exists with the working 
directory (outdir). For instance example 01, scf calculation for Ni 
using CG:

 &control
    calculation='scf'
    restart_mode='from_scratch',
    pseudo_dir = '/home/silviu/Work/O-sesame/pseudo/',
    outdir='/home/silviu/tmp/'
    prefix='ni'
    tprnfor = .true.,
    tstress = .true.
 /
...

output:


     Program PWSCF     v.3.0    starts ...
     Today is  2Oct2005 at 23:28:19

     Ultrasoft (Vanderbilt) Pseudopotentials

     Current dimensions of program pwscf are:

     ntypx = 10   npk = 40000  lmax =  3
     nchix =  6  ndmx =  2000  nbrx = 14  nqfx =  8

 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     from readpp : error #         3
     file /home/silviu/Work/O-sesame/pseudo/NiUS.RRKJ3.UPF not found
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

     stopping ...
~                      

However the file is there:

$ ls  /home/silviu/Work/O-sesame/pseudo/
Al.vbc.UPF           H.fpmd.UPF            NiUS.RRKJ3.UPF      O_US.van
As.gon.UPF           H.vbc.UPF             O.BLYP.UPF          Pb.vdb.UPF
C.meta.tm.UPF        HUSPBE.RRKJ3          O.LDA.US.RRKJ3.UPF  
Ptrel.RRKJ3.UPF
C.pz-rrkjus.UPF      H_US.van              O.pbe-rrkjus.UPF    Si.recon
CVS/                 Hmeta.tm.UPF          O.recon             Si.vbc.UPF
Cu.pz-d-rrkjus.UPF   N.BLYP.UPF            O.vdb.UPF           SiPBE_nc.UPF
Fe.pz-nd-rrkjus.UPF  Ni.pbe-nd-rrkjus.UPF  OPBE_nc.UPF         Ti.vdb.UPF

Any ideas where is the problem?

Thanks, Silviu.





Paolo Giannozzi wrote:

>On Friday 30 September 2005 22:11, Silviu Zilberman wrote:
>
>  
>
>>Thank you for your comments. I had to introduce a couple of other
>>minor changes but now the code compiles. Running however seems
>>still to be a problem, but I hope to make it work.
>>    
>>
>
>Hi Silviu
>
>if you manage to prodce working executables, please let us
>know how you did it. Note that g95 is improving very quickly:
>try the very latest release
>
>Paolo
>
>  
>

-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: make.sys.cygwin_g95
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20051005/5bcfb4cf/attachment.ksh>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: silviu.vcf
Type: text/x-vcard
Size: 247 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20051005/5bcfb4cf/attachment.vcf>