ZG_333.in does not finish

[frederik1]

Hi,

I am running ZG.x for a 3x3x3 CsPbI3 perovskite. There are 135 atoms. The input i used is below:

&input
flfrc='cspbi3.444.fc',
asr='simple',
amass(1)= 132.9054519,
amass(2)= 207.2,
amass(3)= 126.90447,
atm_zg(1) = 'Cs',
atm_zg(2) = 'Pb',
atm_zg(3) = 'I',
flscf = 'cspbi3.scf.in'
T = 0.00,
dim1 = 3, dim2= 3, dim3 = 3
compute_error = .true., synch =.true., error_thresh = 0.05, niters = 30000
incl_qA = .false.
/

The code found an optimum configuration, here is the final lines of the output:

""""""""""
.
.
.
Searching for optimum configuration...
Attempt # 14249
Total error: 0.047571

Optimum configuration found !

Sum of diagonal terms per q-point: 5.729310
Error and niter index: 0.047571 14249

==============================================

Print ZG configuration
""""""""""""""""""

But it has been almost a day (i am running it on a 256GB machine with 64 cpus) and the code is still at the same step. The code is running
while i am writing this message.

So, is this normal? and should i wait for it to complete?

Thank you!

Ilhan Y.

[mzach]

Dear Illhan,

Thank you for posting this question. The problem you are encountering is not normal.

The ZG.x is mostly supposed to run in serial or on a few CPUs apart from some specific cases (https://docs.epw-code.org/doc/TutorialZG.html).

So I suggest you run:

$path_to_QE/bin/ZG.x < ZG_333.in > ZG_333.out

The calculation should take a few seconds/minutes. Let me know if you still have any problems.

Note that you can always increase the error_thesh for speed up; it shouldn't affect the final result dramatically. Any value below 0.2 should be ok.

Best,

Marios Zacharias

[frederik1]

Dear Marios,

Never mind the previous message. When i tried 0.2 again it worked! I can proceed with that.

Thanks a lot!

Best,

Ilhan