[Pw_forum] help on "wrong total_weight"!

[shangy]

Dear all:
   I encountered the error message "wrong total_weight" in my phdos
calculation with serial matdyn.x. The QE-4.0.2 was compiled with
ifort-10.1.015_intal64, lam-7.1.4(and also openmpi-1.2.7) and MKL-10.0.3.020
under RedHat Enterprise Linux Server 5.2 environment. The CPU is Quad-Core
AMD Opteron(tm) Processor 2356. The detail information of my calculation are
as follows:
   I want to obtain the phonon-dos of silicon vacancy (ground state
configuration) with Si63 supercell at gamma point.  The step-by-step
inputfiles are:
for scf calculation
-------------------
 &CONTROL
    calculation  = 'scf'
    verbosity    = 'high'
    restart_mode = 'from_scratch'
    pseudo_dir   = '/home/pub/pplib/'
    outdir = '/tmp/'
    prefix = 'Si-V1'
    tprnfor = .true.
    nstep   = 200
 /
&SYSTEM
    ibrav =  1
    celldm(1) = 20.70
    nat  = 63
    ntyp = 1
    occupations = 'smearing'
    degauss     = 0.002
    smearing    = 'mv'
    ecutwfc = 34
 /
&ELECTRONS
    conv_thr = 1.0D-8
    diagonalization='david'
    mixing_mode  = 'plain'
    mixing_beta = 0.3
 /
ATOMIC_SPECIES
 Si 28.0855 Si.pbe-rrkj.UPF
ATOMIC_POSITIONS (crystal)
.
Si coordinates
.
etc
K_POINTS
1
0.0 0.0 0.0 1.0
-----------------
for phonon calcuation:
----------------------
&INPUTPH
 amass(1) = 28.0855
 outdir = '/tmp/'
 prefix = 'Si-V1'
 tr2_ph = 1.0d-12
 alpha_mix(1) = 0.1
 iverbosity   = 0
 reduce_io = .true.
! recover  = .true.
 fildyn = 'Si-V1.dynG'
! epsil  = .true.
 trans  = .true.
/
0.0  0.0  0.0
--------------

for the IFC calculation:
-----------------------
&input
  fildyn = 'Si-V1.dynG'
  flfrc  = 'Si-V1.fc'
  zasr   = 'crystal'
/
1 1 1
1
Si-V1.dynG
-----------------
All the above calcualtions finished well. However, the forenamed error
message appeared when I calculated the phonon dos with dense q-points.
for the phonon dos calculation:
-------------------------------
&input
  flfrc = 'Si-V1.fc'
  asr = 'crystal'
  dos = .true.
  nk1=10, nk2=10, nk3=10
  deltaE = 0.1
  fldos = 'Si-V1.phdos-10-0.1'
  flfrq = 'Si-V1.freq-10-0.1'
  flvec = 'Si-V1.modes-10-0.1'
  ntyp  = 1
  amass(1) = 28.0855
/
--------------------

The matdyn.x stopped and complained that
--------the matdyn.x outfile-----
     warning: symmetry operation # 34 not allowed.   fractional translation:
      -0.0000387  0.0000000  0.0000387  in crystal coordinates
Norm of the difference between old and new effective charges:
0.00000000000000000000
Norm of the difference between old and new force-constants:
0.01586035815826266834
  0.833333333333333
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     from frc_blk : error #         1
     wrong total_weight
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     stopping ...
---------------------
So, what's wrong with it? code bug or miscompiled?Is something wrong with
the input-parameters in my inputfile? I searched the mailist and found
nothing could solve the problem. BTW, with the similar calculation, I got
optic frequency of 514.4 cm^-1 (expt. 518 cm^-1) at gamma point using the
perfect Si64 suprecell. If you would like to test the problem, the
Si-V1.dynG can be found at http://www.divshare.com/download/5725901-c73 due
to the froum file limits. Please be patient for 15 seconds adversting when
you download it :).
Any ideas are appreciated! Thanks a lot!
Best regards!

shany
Graduate School of Chinese Academy of Sciences
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://www.democritos.it/pipermail/pw_forum/attachments/20081102/eb1da5ec/attachment.htm