[QE-users] [SPAM] Pencil decomposition makes phonon calculations slugish

[jibiaoli]

Dear All,


I try to perform phonon calculations using the command: nohup mpirun -np 48 ph.x <bulk.ph.inp> bulk.ph.out 2>&1&, but got the error below


 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
     Error in routine  fft_type_set (6):
  there are processes with no planes. Use pencil decomposition (-pd .true.)
 %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%


So I tried to using the command: nohup mpirun -np 48 ph.x -pd .true. <bulk.ph.inp> bulk.ph.out 2>&1&, but found that the calculations become extremely slugish, much much slower than fewer cores (24) without -pd .true. Any idea to run ph calculations more efficiently?


Sincerely


Jibiao Li
Sichuan University of Arts and Science



&inputph
search_sym = .false.,
tr2_ph = 1.0d-14,
prefix = 'bulk',
fildvscf = 'aldv',
alpha_mix(1) = 0.15,
amass(1) = 1.00794,
amass(2) = 24.30500,
outdir = './',
fildyn = 'bulk.dyn',
ldisp = .true.,
nq1 = 6, nq2 = 6, nq3 = 6,
/


 &CONTROL
   calculation     = 'scf'
   restart_mode    = 'from_scratch'
   outdir          = './'
   pseudo_dir      = '/home/jibiaoli/pseudo/PAW'
   prefix          = 'bulk'
 /
 &SYSTEM
   ibrav           = 1
   celldm(1)       = 6.398928
   nat             = 25
   ntyp            = 2
   ecutwfc         = 49
   ecutrho         = 411
   occupations     = 'tetrahedra_opt'
 /
 &ELECTRONS
   electron_maxstep = 299
   mixing_beta     = 0.2D0
   diagonalization = 'david'
 /
ATOMIC_SPECIES
    H    1.00790  H.pbe-kjpaw_psl.1.0.0.UPF
   Mg   24.30500  Mg.pbe-spnl-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS angstrom
H             2.9334371019        1.6945880639        2.4064504157
H             2.9361097364        1.6946193258        0.9764573199
H             2.9341317887        2.4103259252        1.6928022975
H             2.9356664900        0.9765087562        1.6928258572
H             0.4563748303        1.6945449253        2.4048843040
H             0.4536753902        1.6945754540        0.9780384352
H             0.4556732876        2.4087508950        1.6927430710
H             0.4541167746        0.9780891614        1.6927663575
H             1.6942074227        2.9338747269        2.4094173440
H             1.6942253553        2.9365230609        0.9734873573
H             2.4084148469        2.9326051884        1.6928151644
H             0.9788601287        2.9338768014        1.6928324483
H             1.6941642845        0.4564001751        2.4044853312
H             1.6941821077        0.4536980304        0.9784450092
H             2.4086627503        0.4549656694        1.6927305862
H             0.9786090540        0.4536705754        1.6927471350
H             1.6942228241        2.4075198302        2.9323288632
H             1.6942361110        0.9758523967        2.9347177575
H             2.4113785197        1.6924498474        2.9330218283
H             0.9759023776        1.6924603570        2.9342779497
H             1.6941592473        2.4044103869        0.4556913284
H             1.6941721664        0.9789728802        0.4532408142
H             2.4082581865        1.6923579184        0.4549779034
H             0.9790135072        1.6923675439        0.4536758309
Mg            0.0000000000        0.0000000000        0.0000000000    0   0   0
K_POINTS automatic
  10 10 10   0 0 0
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20241106/5451809b/attachment.html>