[QE-users] ph.x v6.6 doesn't work on large system

[Candida Pipitone]

Hello,

I can't find a way to make ph.x 6.6 run correctly on my system.
It is a MAPbI3 orthorhombic cell containing 48 atoms (Pb, I, C, N, H), with
a 2x2x4 grid of k-points. ph.x correctly lists all representations (last
line printed is "Representation 144 1 modes - To be done"), it writes the
.wfcXX files and the _ph0 directory in a couple of minutes, then gets stuck
indefinitely (I tested up to one hour). High verbosity does not give
further clues.
A similar system (half the size, 24 atoms) runs fine, starting the
self-consistent calculation right after it lists the representations.
I tried changing the following:
1) crystal setting both specifying the space group and describing the unit
cell with lattice vectors.
2) PPs: PAW/US/full relativistic/PBE/PBEsol
3) the q-grid, or just the gamma point
4) computer: the behavior is the same on a local cluster and Marconi100. In
particular, I ran these tests on M100 with -np 64 and -npool 16
Can you give me any advice on a way to work around this issue, or on some
other tests I could run to help pinpoint what the problem may be?
I attach the pw.x and ph.x input files.
&CONTROL
title ='prova'
calculation ='scf'
restart_mode='from_scratch'
nstep = 200
prefix = 'mapi'
pseudo_dir ='../pseudo/'
outdir ='./'
wf_collect = .TRUE.
/
&SYSTEM
a = 1.26230e+01
c = 6.31612e+00
ibrav = 6
nat = 48
ntyp = 5
ecutwfc = 40
ecutrho = 360
occupations ='smearing'
smearing ='marzari-vanderbilt'
degauss = 0.02
/
&ELECTRONS
conv_thr = 1.0d-6
mixing_beta = 0.7
electron_maxstep = 200
startingpot = 'atomic'
startingwfc = 'atomic+random'
/
ATOMIC_SPECIES
Pb 207.21 Pb.pbe-dn-kjpaw_psl.1.0.0.UPF
I 126.90 I.pbe-n-kjpaw_psl.1.0.0.UPF
C 12.01 C.pbe-n-kjpaw_psl.1.0.0.UPF
N 14.00674 N.pbe-n-kjpaw_psl.0.1.UPF
H 1.00794 H.pbe-kjpaw_psl.0.1.UPF
K_POINTS {automatic}
2 2 4 0 0 0
ATOMIC_POSITIONS {angstrom}
Pb 3.261162 3.037754 3.394201
I 0.105163 3.037754 3.420198
I 3.261162 3.037754 0.181197
C 0.105163 6.193274 5.863317
N 0.105163 6.193274 0.930200
H 0.105163 5.241276 1.267197
H 0.930160 0.357750 1.267197
H 5.591685 0.357750 1.267197
H 5.689684 5.466275 5.499319
H 1.098160 5.927275 5.500317
H 6.150684 0.874751 5.500317
I 3.261162 0.000000 3.394201
Pb 3.261162 9.349274 3.394201
I 0.105163 9.349274 3.420198
I 3.261162 9.349274 0.181197
C 0.105163 12.504794 5.863317
N 0.105163 12.504794 0.930200
H 0.105163 11.552796 1.267197
H 0.930160 6.669270 1.267197
H 5.591685 6.669270 1.267197
H 5.689684 11.777795 5.499319
H 1.098160 12.238795 5.500317
H 6.150684 7.186271 5.500317
I 3.261162 6.311520 3.394201
Pb 9.572682 3.037754 3.394201
I 6.416683 3.037754 3.420198
I 9.572682 3.037754 0.181197
C 6.416683 6.193274 5.863317
N 6.416683 6.193274 0.930200
H 6.416683 5.241276 1.267197
H 7.241680 0.357750 1.267197
H 11.903205 0.357750 1.267197
H 12.001204 5.466275 5.499319
H 7.409680 5.927275 5.500317
H 12.462204 0.874751 5.500317
I 9.572682 0.000000 3.394201
Pb 9.572682 9.349274 3.394201
I 6.416683 9.349274 3.420198
I 9.572682 9.349274 0.181197
C 6.416683 12.504794 5.863317
N 6.416683 12.504794 0.930200
H 6.416683 11.552796 1.267197
H 7.241680 6.669270 1.267197
H 11.903205 6.669270 1.267197
H 12.001204 11.777795 5.499319
H 7.409680 12.238795 5.500317
H 12.462204 7.186271 5.500317
I 9.572682 6.311520 3.394201

&INPUTPH
tr2_ph=1.0d-14,
prefix='mapi'
ldisp = .true.
nq1 = 8,
nq2 = 8,
nq3 = 8,
outdir='./'
fildyn='mapi.dynG',
/

----------------
Candida Pipitone
phD student
University of Palermo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20201105/40a31b3e/attachment.html>