Frequently Asked Questions¶
-
Error in routine bands (1): The bands code with constrained magnetization has not been tested
Error in routine c_bands (1): too many bands are not converged
Error in routine offset_atom_wfc (1): Mismatch between the requested and available manifolds
Error in routine projwave (1): Cannot project on zero atomic wavefunctions!
Error in routine set_occupations (1): smearing requires a vaklue for gaussian broadening (degauss)
Error in routine setup (1): forces for hybrid functionals + US/PAW not implemented
Error in routine system_checkin (12): ntyp too large, increase NSX
General questions, licensing¶
Do I need to pay for Quantum ESPRESSO?¶
Starting with AMS2024, you need to acquire a license from SCM to use Quantum ESPRESSO through AMS. This license also lets you use the SCM-developed periodic DFT code BAND.
Quantum ESPRESSO itself is free and open-source software that is not developed by SCM. See https://www.quantum-espresso.org. You can access the source code, including minor modifications by SCM, on the page Manual compilation and installation.
The benefits of using Quantum ESPRESSO through AMS include:
Easy installation through the AMS package manager on all common platforms
Several pseudopotential libraries are included
Access to all AMS Driver features
Python support through PLAMS
Use the graphical user interface AMSinput to set up calculations
Use the graphical user interfaces AMSmovie (trajectories), AMSview (densities/potentials on grids, isosurfaces, …), AMSbands (electron and phonon band structure) and AMSdos (DOS and PDOS) to analyze results
A single style of input, output, Python support, and GUI common to Quantum ESPRESSO, DFTB, ReaxFF, … which makes it easy to switch between different levels of theory.
Technical (non-scientific) support
The license check happens inside the AMS engine, not inside the externally developed Quantum ESPRESSO code.
Are all Quantum ESPRESSO features supported?¶
No, only some common features as detailed in this manual are supported by the AMS interface.
Programs supported by AMS engine interface (support may not be complete):
bands.x
dos.x
projwfc.x
pw.x
Programs supported by the AMSview GUI module (support may not be complete):
pp.x
Programs not supported by the AMS engine interface nor GUI, but included with the Quantum ESPRESSO package:
abcoeff_to_eps.x , all_currents.x , alpha2f.x , average.x , bands_unfold.x , bse_main.x , casino2upf.x , cell2ibrav.x , cppp.x , cp.x , disca.x , dos_sp.x , dvscf_q2r.x , dynmat.x , ef.x , epa.x , epsilon_Gaus.x , epsilon.x , epw_pp.py , epw.x , ev.x , fd_ef.x , fd_ifc.x , fd.x , fermi_int_0.x , fermi_int_1.x , fermi_proj.x , fermi_velocity.x , fqha.x , fs.x , graph.x , gww_fit.x , gww.x , head.x , hgh2qe.x , hp.x , ibrav2cell.x , initial_state.x , kcwpp_interp.x , kcwpp_sh.x , kcw.x , kpoints.x , lambda.x , ld1.x , manycp.x , matdyn.x , memory_pw4gww.x , molecularnexafs.x , molecularpdos.x , neb.x , open_grid.x , path_interpolation.x , pawplot.x , phcg.x , ph.x , plan_avg.x , plotband.x , plotproj.x , plotrho.x , pmw.x , postahc.x , postw90.x , ppacf.x , pp_disca.x , pprism.x , pp_spctrlfn.x , pw2bgw.x , pw2critic.x , pw2gt.x , pw2gw.x , pw2wannier90.x , pw4gww.x , pwcond.x , pwi2xsf.x , q2qstar.x , q2r.x , qe_xc_infos.x , scan_ibrav.x , simple_bse.x , simple_ip.x , simple.x , spectra_correction.x , sumpdos.x , turbo_davidson.x , turbo_eels.x , turbo_lanczos.x , turbo_magnon.x , turbo_spectrum.x , upfconv.x , virtual_v2.x , w90chk2chk.x , wannier90.x , wannier_ham.x , wannier_plot.x , wfck2r.x , wfdd.x , xspectra.x , ZG.x
How do I convert my old Quantum ESPRESSO inputs to the new AMS2024 format?¶
AMSinput attempts to convert some of the input options, but there is no guarantee that it does so correctly. Double-check the input options carefully.
If you prefer to use the GUI to standalone Quantum ESPRESSO, you need to use AMS2018-AMS2023.
Can I run Quantum ESPRESSO with GPU acceleration?¶
The Quantum ESPRESSO package that is included with AMS does not have any GPU support.
You may be able to compile Quantum ESPRESSO yourself such that it both runs on the GPU and supports the AMS interface.
Where is Quantum ESPRESSO installed? Where can I find pw.x etc.?¶
First make sure that you have installed Quantum ESPRESSO through the AMS package manager.
Run this command:
"$AMSBIN/amspackages" loc qe
See the AMSpackages documentation for more details.
If you uninstall AMS, you can also delete this directory.
How do I run standalone Quantum ESPRESSO without going through the AMS interface?¶
To run Quantum ESPRESSO as a standalone program without going through AMS, you
need to set the correct environment variables with the $AMSBIN/startqe
shell
script (note the leading dot):
. "$AMSBIN/startqe"
Alternatively, you can just launch Quantum ESPRESSO as follows:
"$AMSBIN/startqe" pw.x
In a similar way, you can launch ph.x
, pp.x
, etc.
See also
How do I calculate the band structure with hybrid functionals?¶
The AMS interface to Quantum ESPRESSO does not support this.
Instead, use BAND, see for example this tutorial for calculating effective masses.
How do I use functionals from libxc?¶
The QE package that is installable from the AMS package manager is not linked against libxc. It is therefore not possible to use functionals from libxc.
You may be able to compile QE yourself and couple it to the AMS Driver.
The program is always running in serial (using 1 core) ?¶
When running Quantum ESPRESSO through AMS, you need to set the environment
variable SCM_DISABLE_MPI=1
. This causes AMS to launch in serial
and you will see a line like this at the top:
AMS 2024.101 RunTime: Apr10-2024 06:40:44 ShM Nodes: 1 Procs: 1
This is expected and intended. This line does not mean that Quantum ESPRESSO is running in serial.
Further down in the standard output, at least for single-point calculations, you will find information about how QE is parallelized:
Starting QuantumEspresso worker ...
Command: "$AMSBIN/startqe" pw.x -in "/path/qeworker_1124704473.qein" --amspipe
Program PWSCF v.7.1 starts on 10Apr2024 at 6:40:46
Git branch: HEAD
Last git commit: af50f724259d49e570aeba1e2adb5da6c31b8e69
Last git commit date: Fri Mar 15 14:41:38 2024 +0100
Last git commit subject: Link libamspipe statically on Windows
This program is part of the open-source Quantum ESPRESSO suite
for quantum simulation of materials; please cite
"P. Giannozzi et al., J. Phys.:Condens. Matter 21 395502 (2009);
"P. Giannozzi et al., J. Phys.:Condens. Matter 29 465901 (2017);
"P. Giannozzi et al., J. Chem. Phys. 152 154105 (2020);
URL http://www.quantum-espresso.org",
in publications or presentations arising from this work. More details at
http://www.quantum-espresso.org/quote
Parallel version (MPI & OpenMP), running on 8 processor cores
Number of MPI processes: 8
Threads/MPI process: 1
MPI processes distributed on 1 nodes
2022 MiB available memory on the printing compute node when the environment starts
In this case, we see that QE is actually running on 8 processor cores.
You can also monitor the CPU usage using top
or other utilities.
Warnings and notes¶
Windows Security: Do you want to allow public and private networks to access this app?¶
“Windows Firewall has blocked some features of pw.x.exe on all public and private networks”.
This warning can appear when launching pw.x (or bands.x, dos.x, etc.) for the first time on Windows.
This is a spurious warning that has no effect when running QE on a single machine. You do not need to allow any network access, and can just click “Cancel” in the dialog window.
Note: a single QE.Label is assigned to atoms of different species¶
This message can appear when you have regions. For example, if one C atoms is in region1 and another C atom is in region2.
The message can safely be ignored.
using ibrav=0 with symmetry is DISCOURAGED, use correct ibrav instead¶
This message appears in all calculations using the AMS interface to Quantum ESPRESSO. The AMS engine sets ibrav=0.
The message can be ignored. It is not possible to set ibrav to nonzero values through the AMS interface.
When it is extra important to have a continuous potential energy surface (for example for the numerical calculation of phonons), AMS will automatically disable the symmetry detection in QE, so that the symmetric and non-symmetric points on the PES have the exact same k-point grids etc.
SCF convergence tips¶
Decrease Electrons%mixing_beta¶
You can try to decrease Electrons%mixing_beta
, or vary other settings in the Electrons
block (on the Details → SCF panel in AMSinput).
Errors¶
Error in routine bands (1): The bands code with constrained magnetization has not been tested¶
Run the SCF calculation without constrained magnetization before calculating the band structure.
Error in routine c_bands (1): too many bands are not converged¶
Error in routine offset_atom_wfc (1): Mismatch between the requested and available manifolds¶
This message can appear when you run Hubbard DFT+U calculations, for example with the mt_fhi pseudopotentials. Try a different set of pseudopotentials.
You may also be able to manually modify the pseudopotential files to contain the correct information. See the Quantum ESPRESSO documentation and mailing list for details.
Error in routine projwave (1): Cannot project on zero atomic wavefunctions!¶
This error might appear if you try to calculate the PDOS. Do not use ONCV pseudopotentials (e.g. from the SG15 family of pseudopotentials)
Error in routine set_occupations (1): smearing requires a vaklue for gaussian broadening (degauss)¶
You have to set a smearing larger than 0.
Error in routine setup (1): forces for hybrid functionals + US/PAW not implemented¶
Use norm-conserving pseudopotentials, for example from the PseudoDojo or SG15 library.
Error in routine sym_rho_init_shell (2): lone vector¶
You may get this error at the end of lattice optimizations with norm-conserving pseudopotentials (e.g., SG15, Dojo) if the energy cutoff is not large enough. This might result in in unrealistic structures. Increase the energy cutoff (and make sure ecutrho
= 4*ecutwfc
, which is the default). Note that you may need to increase ecutwfc
to values higher than the minimum recommended values in this table
Error in routine system_checkin (12): ntyp too large, increase NSX¶
You may only have up to 10 atom types (chemical elements or unique values for QE.Label
) in the system.
ERROR: ‘EngineRestart’ option to restart a calculation from a previous wave function requires the .save directory¶
You may get this error if
the
quantumespresso.save
directory does not exist in the same directory as thequantumespresso.rkf
file used for the EngineRestart, oryou try to run the EngineRestart calculation on a different machine (for example, a different compute cluster) compared to the original calculation. The graphical user interface does not copy the
quantumespresso.save
directory, only thequantumespresso.rkf
file to the remote machine.
I get a different error¶
If the error is reported by Quantum ESPRESSO, see the Quantum ESPRESSO troubleshooting guide. Errors reported by Quantum ESPRESSO are typically prefixed by “QuantumEspresso” in the .log file and in AMSjobs:
You can also ask questions on the Quantum ESPRESSO mailing list, but please note that
You should not attach the AMS .run/.in/.ams file, but instead
Find the file ending in
.qein
in the results directory,Reproduce the error by (1) copying the
filename.qein
file into an empty directory or a directory containing all needed input files, and (2) running$AMSBIN/startqe pw.x < filename.qein
. (or replacepw.x
withdos.x
,bands.x
, etc., depending on which program should be executed)Ask your question on the mailing list using the input/output from the step above
If you receive any help, you will need to try to convert the information into AMS input yourself
If the error is reported by the AMS Driver, see the AMS Driver documentation