[Pw_forum] Example 12

[W2AGZ]

To All...

 

Running example 12 from QE 4.2.1 within an Ubuntu 10.04.1 distro results in
the following abend messages:

 

>From the terminal...

............

running pwcond.x to calculate transmission of a perfect Al wire ...STOP 2

Error condition encountered during test: exit status = 2

Aborting

............

 

>From AlWireAl.cond.out......

...........

E-Ef(ev), T(x2 spins) =    1.5500000   6.0000000

 

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%

     from gep_x : error #        38

     error on zggev

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%

 

     stopping ...

.........

I vaguely remember problems with zggev being reported in the Fora one or two
years ago.  Now, running QE 4.2.1 example 12 under an Ubuntu 8.04.4 umbrella
goes just fine.  So...I smell a fortran compiler problem here.  8.04.4 comes
with gfortran-4.2, whilst 10.04.1 has "upgraded" to gfortran-4.4.  Anyone
else run into gfortran problems with QE 4.2.1?  A few months ago, Eyvaz
solved a gfortran problem I was having with lambda.f90.

 

On the general issue of fortran compilers, I've noticed over the years
whenever a discussion ensues between advocates (zealots) of one particular
edition over another, the collective IQ can descend to 30.  I've heard both
good and bad things about gfortran, but there does seem to be some agreement
that it produces reasonably optimized code.  And it's free...cheap at twice
the price.

 

Paul Grant

www.w2agz.com

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20101113/e727b6bc/attachment.html>