[Pw_forum] << running QE in parallel >>

[I. Camps]

SOLVED!

The problem is the script run_example.
If I run the input file directly from the prompt using

$PARA_PREFIX pw.x $PARA_POSTFIX < al001.rx.in > al001.rx.out

I get the following output:

***************************
     Program PWSCF     v.4.1    starts ...
     Today is 28Aug2009 at 12:34:42

     Parallel version (MPI)

     Number of processors in use:       4
     R & G space division:  proc/pool =    4

     For Norm-Conserving or Ultrasoft (Vanderbilt) Pseudopotentials or PAW

     Current dimensions of program pwscf are:
     Max number of different atomic species (ntypx) = 10
     Max number of k-points (npk) =  40000
     Max angular momentum in pseudopotentials (lmaxx) =  3
     Waiting for input...

     Subspace diagonalization in iterative solution of the eigenvalue problem:
     a parallel distributed memory algorithm will be used,
     eigenstates matrixes will be distributed block like on
     ortho sub-group =    2*   2 procs

........
***************************

[]s,

Camps


On Fri, Aug 28, 2009 at 11:42 AM, I. Camps<i_camps at yahoo.com.br> wrote:
> Hello all,
>
> I have a question about running QE in parallel.
>
> I installed LAM/MPI and then compiled QE with the parallel default options.
>
> Then I started running the examples. In the output of example 03, I
> get the following:
>
> ***************************
>  Program PWSCF     v.4.1    starts ...
>      Today is 28Aug2009 at 11:30:28
>
>      Parallel version (MPI)
>
>      Number of processors in use:       1
>
>      For Norm-Conserving or Ultrasoft (Vanderbilt) Pseudopotentials or PAW
>
>      Current dimensions of program pwscf are:
>      Max number of different atomic species (ntypx) = 10
>      Max number of k-points (npk) =  40000
>      Max angular momentum in pseudopotentials (lmaxx) =  3
>      Waiting for input...
>      file O.pz-rrkjus.UPF: wavefunction(s)  2S renormalized
>      file C.pz-rrkjus.UPF: wavefunction(s)  2S renormalized
>
>      gamma-point specific algorithms are used
>
>
>      Subspace diagonalization in iterative solution of the eigenvalue problem:
>      Too few procs for parallel algorithm
>        we need at least 4 procs per pool
>      a serial algorithm will be used
>
> ..................
> ***************************
>
> So as you can see, pw.x is running in serial mode.
>
> How do I solve this issue?
>
> My system:
> Processor: AMD Phenom(tm) II X4 940 @ 3GHz (Deneb)
> Memory: 8GB RAM
> Lam version (from laminfo): 7.1.4
> OS: Ubuntu 9.04 64bits (up-to-date)
> Environment variables:
>
> export TOPDIR='/software/espresso-4.1'
> export BIN_DIR=$TOPDIR/bin
> export PSEUDO_DIR=$TOPDIR/pseudo
> export TMP_DIR=/scratch
> export PARA_PREFIX="mpiexec -n 4"
> export PARA_POSTFIX="-npool 1"
> PWgui="/software/PWgui-4.1"
>
>
> Regards,
>
> Camps
>