[Pw_forum] PWscf calculation problem （mpi and serial）

[Jiaye, Li]

Dear users

I installed pwscf on the RH.Linux with 2G memory(2 CPUs and 4 processes).
MPICH1.2.7,Ifort,lapack-3.1.1, BLAS and fftw-3.2 have been successfully
installed, too. The subsequent installation process of PWscf report no
error. All the example tests are successful. In addition, I selected a
surface slab model to make another couples of test. I met some problems:
1. I typed " mpirun -np 4 ~/bin/pw.x <IO.inp |tee IO.out &" , then I
followed the content of output file closely, finding it stops after the
lines:

     number of k points=    4  gaussian broad. (Ry)=  0.0200     ngauss =
0
                       cart. coord. in units 2pi/a_0
        k(    1) = (   0.0000000   0.0000000   0.0000000), wk =   0.5000000
        k(    2) = (   0.0000000   0.0000000  -1.0872780), wk =   0.5000000
        k(    3) = (   0.0000000  -1.0872780  -1.0872780), wk =   0.5000000
        k(    4) = (   0.0000000  -1.0872780   0.0000000), wk =   0.5000000
     G cutoff =17198.9000  (1997991 G-vectors)     FFT grid: (270,125,125)
     G cutoff = 2293.1867  (  97139 G-vectors)  smooth grid: ( 96, 45, 45)
     Largest allocated arrays     est. size (Mb)     dimensions
        Kohn-Sham Wavefunctions        14.87 Mb     (   3074, 317)
        NL pseudopotentials            27.02 Mb     (   3074, 576)
        Each V/rho on FFT grid         16.48 Mb     (1080000)
        Each G-vector array             3.81 Mb     ( 499499)
        G-vector shells                 0.70 Mb     (  91837)
     Largest temporary arrays     est. size (Mb)     dimensions
        Auxiliary wavefunctions        59.48 Mb     (   3074,1268)
        Each subspace H/S matrix       24.53 Mb     (   1268,1268)
        Each <psi_i|beta_j> matrix      2.79 Mb     (    576, 317)
        Arrays for rho mixing         131.84 Mb     (1080000,   8)

but pw.x process is still occupying 4x25 % cpu and 10 % MEM. I wait for
hours and there are no further new contents read from the output file. I
don't know whether it is the overflow of the memery? Could anybody help me
with these problems? Thanks.

Attached file is the make.sys of the PWscf.



-- 
Sincerely

Jiaye Li



==============================================
Li, Jiaye

M.S. in designing and developing novel materials

Research experience:
*ALD precursor. ie. Metal complexes
*Surface modification. ie. Indium tin oxide, SiO2
*Gas storage and adsorption. ie. Ionic Liquids

Skills:
*ab-initial method, DFT simulation and Force Field simulation

Please contact me at:
1. jameslipd at gmail.com  or
2. jameslibd at gmail.com

==============================================
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20081222/30397227/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: make.zip
Type: application/zip
Size: 2095 bytes
Desc: not available
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20081222/30397227/attachment.zip>