I am afraid that:
high cutoff + many k-points + hybrid functionals + PAW = an enormous
amount of time
Paolo
On 10/04/2024 14:08, aleksandr.doma--- via users wrote:
Hi!
I did dft calculations with the hybrid HSE functional. The mail part of
the input file is here:
&control
prefix = 'lkbo',
pseudo_dir='./'
outdir = './tmp'
calculation = "scf"
/
&system
A = 8.49150
B = 11.1415
C = 12.6558
ecutrho = 480
ecutwfc = 60
ibrav = 8
nat = 104
ntyp = 4
nbnd = 320
occupations = 'tetrahedra'
input_dft='hse', nqx1 = 3, nqx2 = 2, nqx3 = 2,
/
&electrons
mixing_beta = 0.5
/
ATOMIC_SPECIES
B3+ 10.8110 B.pbe-n-kjpaw_psl.0.1.UPF
K1+ 39.0983 K.pbe-spn-kjpaw_psl.1.0.0.UPF
O2- 15.9994 O.pbe-n-kjpaw_psl.1.0.0.UPF
Li 6.9410 Li.pbe-sl-kjpaw_psl.1.0.0.UPF
ATOMIC_POSITIONS {angstrom}
K1+ 1.634019 1.375641 2.513315
K1+ 2.611731 9.765859 8.841215
K1+ 5.879769 4.195109 10.142485
................................................................
................................................................
K_POINTS {automatic}
6 4 4 0 0 0
In order to provide the necessary resources (primarily RAM), I set a
walltime of 12 hours (the maximum available). During this time, only
self-consistent calculations were made.... the output file ended with this:
k =-0.5000-0.3811-0.3355 ( 63480 PWs) bands (ev):
-40.2074 -40.2074 -40.2074 -40.2074 -39.8136 -39.8136 -39.8136 -39.8136
-25.0246 -25.0246 -25.0246 -25.0246 -24.7741 -24.7741 -24.7741 -24.7741
-17.7752 -17.7752 -17.7752 -17.7752 -17.4711 -17.4711 -17.4711 -17.4711
-17.1362 -17.1362 -17.1362 -17.1362 -16.5958 -16.5958 -16.5958 -16.5958
-16.2415 -16.2415 -16.2415 -16.2415 -16.0495 -16.0495 -16.0495 -16.0495
-15.8145 -15.8145 -15.8145 -15.8145 -15.6181 -15.6181 -15.6181 -15.6181
-15.4151 -15.4151 -15.4151 -15.4151 -15.2465 -15.2465 -15.2465 -15.2465
-15.2143 -15.2143 -15.2143 -15.2143 -15.1745 -15.1745 -15.1745 -15.1745
-14.7186 -14.7186 -14.7186 -14.7186 -14.0298 -14.0298 -14.0298 -14.0298
-8.5758 -8.5758 -8.5758 -8.5758 -8.5494 -8.5494 -8.5494 -8.5494
-8.4877 -8.4877 -8.4877 -8.4877 -8.3338 -8.3338 -8.3338 -8.3338
-8.3044 -8.3044 -8.3044 -8.3044 -8.2670 -8.2670 -8.2670 -8.2670
-6.0841 -6.0841 -6.0841 -6.0841 -5.8829 -5.8829 -5.8829 -5.8829
-5.6034 -5.6034 -5.6034 -5.6034 -5.1035 -5.1035 -5.1034 -5.1034
-4.6032 -4.6032 -4.6032 -4.6032 -4.2454 -4.2454 -4.2454 -4.2454
-4.0476 -4.0476 -4.0476 -4.0476 -3.8141 -3.8141 -3.8141 -3.8141
-3.5191 -3.5191 -3.5191 -3.5191 -3.1688 -3.1688 -3.1688 -3.1688
-2.8103 -2.8103 -2.8103 -2.8103 -2.5451 -2.5451 -2.5451 -2.5451
-2.3255 -2.3255 -2.3255 -2.3255 -2.0994 -2.0994 -2.0994 -2.0994
-2.0127 -2.0127 -2.0127 -2.0127 -1.8472 -1.8472 -1.8472 -1.8472
-1.6640 -1.6640 -1.6640 -1.6640 -1.2849 -1.2849 -1.2849 -1.2849
-0.9584 -0.9584 -0.9584 -0.9584 -0.5666 -0.5666 -0.5666 -0.5666
-0.4752 -0.4752 -0.4752 -0.4752 -0.1894 -0.1894 -0.1894 -0.1894
-0.0546 -0.0546 -0.0546 -0.0546 0.4020 0.4020 0.4021 0.4021
0.5736 0.5736 0.5736 0.5736 0.6209 0.6209 0.6209 0.6209
0.7237 0.7237 0.7237 0.7237 0.8574 0.8574 0.8574 0.8574
0.9494 0.9494 0.9494 0.9494 1.0333 1.0333 1.0333 1.0333
1.1146 1.1146 1.1146 1.1146 1.1520 1.1520 1.1520 1.1520
1.2932 1.2932 1.2932 1.2932 1.3902 1.3902 1.3902 1.3902
1.4360 1.4360 1.4360 1.4360 1.5876 1.5876 1.5876 1.5876
1.6916 1.6916 1.6916 1.6916 1.7709 1.7709 1.7709 1.7709
1.9443 1.9443 1.9443 1.9443 2.1543 2.1543 2.1543 2.1543
2.4438 2.4438 2.4438 2.4438 2.4933 2.4933 2.4933 2.4933
8.8739 8.8739 8.8739 8.8739 9.0619 9.0619 9.0619 9.0619
9.2578 9.2578 9.2578 9.2578 9.6131 9.6131 9.6131 9.6131
9.8197 9.8197 9.8197 9.8197 10.0460 10.0460 10.0460 10.0460
10.5061 10.5061 10.5061 10.5061 11.0164 11.0164 11.0164 11.0164
11.2186 11.2187 11.2187 11.2187 11.4402 11.4402 11.4402 11.4402
11.6986 11.6986 11.6986 11.6986 11.8813 11.8813 11.8813 11.8813
12.1292 12.1292 12.1293 12.1294 12.3604 12.3614 12.3633 12.3666
the Fermi energy is 2.5553 ev
! total energy = -3569.17632732 Ry
estimated scf accuracy < 0.00000030 Ry
convergence has been achieved in 18 iterations
Using ACE for calculation of exact exchange
EXX grid: 507361 G-vectors FFT dimensions: ( 80, 108, 120)
ACE projected onto 320 (nbndproj) and applied to 320 (nbnd) bands
forrtl: error (78): process killed (SIGTERM)
Image PC Routine Line Source
libpthread-2.28.s 00001540000BDCF0 Unknown Unknown Unknown
Please write if possible:
1) use the obtained result (for example, charge-density.dat) to run the
same calculation? How to do this and will it help to perform the
calculation with the same resources (140 cores, 5 TB RAM)?
2) simplify the input file to speed up the calculation (or reduce the
required resources)
Thanks in advance,
Alex
_______________________________________________
The Quantum ESPRESSO community stands by the Ukrainian
people and expresses its concerns about the devastating
effects that the Russian military offensive has on their
country and on the free and peaceful scientific, cultural,
and economic cooperation amongst peoples
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
--
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 206, 33100 Udine Italy, +39-0432-558216
_______________________________________________
The Quantum ESPRESSO community stands by the Ukrainian
people and expresses its concerns about the devastating
effects that the Russian military offensive has on their
country and on the free and peaceful scientific, cultural,
and economic cooperation amongst peoples
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
users mailing list users@lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
