[Pw_forum] HSE calculation
Giuseppe Mattioli
giuseppe.mattioli at ism.cnr.it
Thu Jan 5 14:30:51 CET 2017
Dear Jiqiang Li
EXX calculations *are* expensive. First of all read this file:
your-path-to-QE/Examples/PW/EXX_example/README
This said, you can use a couple of speedup tricks:
a) The Adaptively Compressed Exchange (ACE) algorithm is implemented
in QE 6.0. EXX is computed with ACE if you add a -D__EXX_ACE flag to
DFLAGS in the make.inc file and recompile. You can obtain a x3-x4
speedup in your EXX calculations.
b) ecutfock is a very useful handle that reduces the ecutrho fft grid
in the case of exact exchange calculations. You should play around to
find convergence, but in my experience ecutrho=ecutwfc yields
satisfactory geometry optimizations, and ecutrho=2*ecutwfc yields
converged eigenvalues and energies.
c) adaptive_thr=.true. is very useful to reduce the number of
iterations in "inner" scf loops, unless your system is very hard to
converge.
Note also that:
d) K_POINTS gamma. This is a good starting point, but keep in mind
that 'nscf' or 'bands' calculations are not allowed in the case of
EXX. You should wonder therefore whether you need a proper k-points
mesh (that of course will slow down your calculation...). The same
holds for nqx1, nqx2, nqx3: in principle you should check the
convergence of the default =1 values...
e) conv_thr = 1.0D-10. Too tight. This will also increase the number
of 'inner' scf cycles without reason. 1.0D-7 is good for energies and
eigenvalues. In 'relax' calculations you can start with 1.0D-6 and the
value will be adaptively lowered to 1.0D-8 when you approach
convergence on forces.
f) nosym = .true. Use symmetry if there is any. EXX calculations will
be faster.
g) Cu.nc.pbe.UPF, ecutwfc = 40. If "nc" means "norm conserving", then
40 Ry is likely a poorly converged value. Check PP convergence on
simpler systems, but I expect no less than 80 Ry, and maybe more...
h) 7.540540000000000E-005. 0 is 0! :-) Try to use values that help the
code to find symmetry (with a confidence of 1.0D-6, if I remember
well), as this will speedup the calculation.
i) disk_io = 'none'. Dangerous. It could hinder a proper restart of
your calculation.
j) degauss=0.0015. Too tight. Start with 0.01 (the value is in Ry -->
0.136 eV), check the smearing contribution to total energy and if it
is large then try to lower the value.
HTH
Giuseppe
Quoting jqli14 at fudan.edu.cn:
> Dear Quantum Espresso users,
> When I did a HSE calculation, I found it was very expensive. I
> am not sure whether I did it in the right way, and could anyone give
> me some advice on the speeding up the calculation?
> The following is my input file. Thank you for your consideration!
>
> Best wishes!
> Jiqiang Li
> Fudan University, China
> %%%%%%%%%%%%%%%%%%%%%%%%%%%
> &CONTROL
> title = '1'
> calculation = 'scf'
> restart_mode = 'from_scratch'
> outdir = './1'
> pseudo_dir = './'
> prefix = '1'
> tstress = .true.
> tprnfor = .true.
> disk_io = 'none'
> verbosity = 'high'
>
> /
> &SYSTEM
> ibrav = 0
> nat = 64
> ntyp = 4
> ecutwfc = 40
> nosym = .true.
> tot_charge = 2
> occupations='smearing', smearing='gaussian', degauss=0.0015
>
> input_dft = 'hse'
> nqx1 = 1
> nqx2 = 1
> nqx3 = 1
> exx_fraction = 0.25
>
> /
> &ELECTRONS
> diagonalization = 'david'
> conv_thr = 1.0D-10
> ! mixing_beta = 0.3
> /
> !&IONS
> ! ion_dynamics = 'bfgs'
> ! /
>
> ATOMIC_SPECIES
> Cu 63.546 Cu.nc.pbe.UPF
> Zn 65.38 Zn.nc.pbe.UPF
> Sn 118.710 Sn.nc.pbe.UPF
> S 32.06 S.nc.pbe.UPF
>
>
> K_POINTS gamma
>
>
>
>
>
>
>
>
> CELL_PARAMETERS angstrom
> 10.8901996613000 0.000000000000000E+000 0.000000000000000E+000
> 7.540540000000000E-005 10.8900995252000 0.000000000000000E+000
> -1.330617000000000E-004 -6.450591000000001E-004 10.8617000380000
>
> ATOMIC_POSITIONS crystal
> Cu 0.501822825617239 8.209585049845000E-004 -3.587372715888000E-004
> Cu 0.247123584741201 0.245878105489351 0.497638837593854
> Cu -3.300201265132100E-003 0.246297973190239 0.747736115954145
> Cu 0.249999956404917 -9.075512491000000E-007 0.252075935894797
> Cu 0.998167806938834 -8.247521468438000E-004 -3.614930078269000E-004
> Cu 0.750818514719231 0.248169430879793 0.500387215747760
> Cu 0.503702437589932 0.246700288309942 0.752294883959966
> Cu 0.750000142082280 -5.261305244000000E-007 0.250013471551898
> Cu 0.504110794881591 0.497130019670560 2.387199659440000E-003
> Cu 0.252864021824887 0.754114287465916 0.497641282622189
> Cu -3.704195077655900E-003 0.753301366753437 0.752296142763722
> Cu 0.249999236820237 0.499998914444942 0.250013182582039
> Cu 0.995875748692389 0.502868653338768 2.390474663852800E-003
> Cu 0.749173286025747 0.751824906786787 0.500389207464035
> Cu 0.503301976072833 0.753704886063385 0.747732698188728
> Cu 0.750000264340525 0.499999872289094 0.247953552793443
> S 0.121607711130560 0.119066726107911 0.128004116420937
> S 0.377410970694754 0.368337305224885 0.623452478431281
> S 0.126885481660098 0.128142136277065 0.628104013094463
> S 0.377653504065843 0.379487027813697 0.128526410123051
> S 0.372131518633960 0.127780399961450 0.372629646011111
> S 0.118342036620952 0.372594923808515 0.876575437772911
> S 0.129491327114066 0.372352660069188 0.371503083918088
> S 0.379334214207609 0.122123268492755 0.871079119116580
> S 0.622980711218032 0.121230496230355 0.128057562219561
> S 0.872125990350811 0.370669474457333 0.628947185465573
> S 0.627135346433278 0.128362738539397 0.627319397985374
> S 0.877782158592530 0.377872619579312 0.127400953525378
> S 0.871236086814705 0.127025188823458 0.371968945581645
> S 0.621861966674097 0.376885089946700 0.871927004252818
> S 0.630936881905509 0.371604494475657 0.372021999310852
> S 0.878366210795192 0.122870165561448 0.872709760521635
> S 0.122354911128635 0.620512219317389 0.128523510696737
> S 0.373121190301768 0.871857201740837 0.628100325702825
> S 0.122595407121820 0.631661221636976 0.623457680920035
> S 0.378401629960358 0.880931755370115 0.128007691557354
> S 0.370516759927977 0.627650572115622 0.371503234800683
> S 0.120676019249323 0.877880018784136 0.871080486353522
> S 0.127877053082689 0.872222465737074 0.372628362873367
> S 0.381665109983142 0.627410813576234 0.876573570642188
> S 0.622225243791753 0.622126275574244 0.127398395493092
> S 0.872872581014923 0.871637376796884 0.627316989829439
> S 0.627882496273248 0.629329558823215 0.628949791420298
> S 0.877028056968618 0.878769311284339 0.128059546753167
> S 0.869072985188352 0.628398420306380 0.372023399476006
> S 0.621642750204393 0.877133981256401 0.872711115745701
> S 0.628774116539949 0.872978079000506 0.371969741830591
> S 0.878145790717181 0.623118862812509 0.871926456963634
> Sn -7.608358306536000E-004 3.222581925182000E-004 0.499443342221132
> Sn 0.256283989515477 0.242473334236634 7.086312024705200E-003
> Sn 0.250000147719144 0.499998628230229 0.750015148413511
> Sn 0.500756332281728 -3.228649230890000E-004 0.499443668263479
> Sn 0.749674703021384 0.249241762433395 5.853144321951000E-004
> Sn -7.528252332882800E-003 0.493714016457111 0.492943834309768
> Sn 0.243712085123441 0.757526932064565 7.088098855306700E-003
> Sn 0.507524071793615 0.506285644594180 0.492942589729261
> Sn 0.750319939734500 0.750758314128771 5.864075895717000E-004
> Zn -8.444909379011000E-004 0.248644383328342 0.249824392514012
> Zn 0.501355703397740 0.249153332059540 0.250202955214345
> Zn 0.750000814591615 0.499998819958268 0.751230280703815
> Zn -1.354273739554800E-003 0.750845148370982 0.250202011729470
> Zn 0.250001030302999 0.999999919970564 0.748797640094973
> Zn 0.500847542910512 0.751353999698878 0.249827089512433
> Zn 0.750000701673651 6.633853790000001E-008 0.750013712421637
--
********************************************************
- Article premier - Les hommes naissent et demeurent
libres et égaux en droits. Les distinctions sociales
ne peuvent être fondées que sur l'utilité commune
- Article 2 - Le but de toute association politique
est la conservation des droits naturels et
imprescriptibles de l'homme. Ces droits sont la liberté,
la propriété, la sûreté et la résistance à l'oppression.
********************************************************
Giuseppe Mattioli
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
v. Salaria Km 29,300 - C.P. 10
I 00015 - Monterotondo Stazione (RM)
Tel + 39 06 90672836 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>
http://www.ism.cnr.it/en/staff/giuseppe-mattioli/
ResearcherID: F-6308-2012
More information about the users
mailing list