[QE-users] Memory issues using pw.x

[Giuseppe Mattioli]

Dear Joseph
First thing first, as you are using norm-conserving pseudopotentials  
(and this is good for hybrid exact-exchange functionals), you don't  
need to specify ecutrho that is fixed at ecutwfc*4. You should specify  
ecutfock instead, ranging from ecutfock=ecutwfc (good enough for  
geometry optimizations) and ecutfock=2*ecutwfc (accurate for  
properties). This saves a lot of computational resources. It doesn't  
matter that the GGA (i.e., non hybrid) calculation works, nor that the  
first scf step is performed (the first step is a GGA one). The problem  
may arise when the code makes the first attempt to calculate EXX with  
the ACE algorithm.
HTH
Giuseppe

Quoting Joe Page <J.Page3 at lboro.ac.uk>:

> Dear Giuseppe,
>
> Here is the input file I want to run:
> &CONTROL
>                 title = 'In2Se3'
>                 calculation = 'scf',
>                 verbosity = 'high',
>                 outdir = './',
>                 pseudo_dir = '/home/ph/phjp2/Psuedopotential_files',
>                 prefix = 'In2Se3_beta3R_Bulk_PBE_SOC_PAW_Fixed_ibrav_5_out',
>                 disk_io = 'high',
>                 restart_mode = 'from_scratch'
> /
> &SYSTEM
>                 ibrav = 5,
> celldm(1) =     18.61578036
> celldm(4) =      0.91627911
>     nat = 5
>     ntyp = 2
>                 ecutwfc = 70,
>                 ecutrho = 600,
>                 occupations = 'fixed',
>                 vdw_corr = 'grimme-d3'
>                 input_dft = "b3lpv1r",
>                 noncolin = .true
>                 lspinorb = .true
>                 nbnd = 60
> /
> &ELECTRONS
>                 conv_thr = 1.D-8
> /
> ATOMIC_SPECIES
> Se 34.00 Se_ONCV_PBE-1.1.upf
> In 49.00 In_ONCV_PBE-1.1.upf
> K_POINTS automatic
> 7 7 5 0 0 0
> ATOMIC_POSITIONS { angstrom }
> Se           -0.0993938971        0.1012574621        6.1581205918
> In            1.9470115521       -1.0456897715       26.6217360387
> Se            0.2247971585       -0.2376878968       -0.1655724739
> In           -0.1239292111       -2.2289160682        1.7398283120
> Se            1.9356313130       -1.0739556190        3.0694920622
>
> And here is the estimated RAM required, from the .out file:
>
>      Estimated static dynamical RAM per process >     680.35 MB
>
>      Estimated max dynamical RAM per process >     681.61 MB
>
>      Estimated total dynamical RAM >      79.88 GB
>
> The pw.x file that is working has similar number to the above parameters.
>
> Hope this helps, and if there is anymore you need just let me know.
>
> Thank you,
> Joe
>
> From: users <users-bounces at lists.quantum-espresso.org> on behalf of  
> Giuseppe Mattioli <giuseppe.mattioli at ism.cnr.it>
> Date: Friday, 20 January 2023 at 11:23
> To: Quantum ESPRESSO users Forum <users at lists.quantum-espresso.org>
> Subject: Re: [QE-users] Memory issues using pw.x
> ** THIS MESSAGE ORIGINATED OUTSIDE LOUGHBOROUGH UNIVERSITY **
>
> ** Be wary of links or attachments, especially if the email is  
> unsolicited or you don't recognise the sender's email address. **
>
> Dear Joseph
> Impossible to say more without seeing your input and knowing your
> available resources.
> Is the estimate printed close to the beginning of the pw.x output, like this
>
>       Estimated max dynamical RAM per process >       1.51 GB
>
>       Estimated total dynamical RAM >      48.27 GB
>
> compatible with the available resources you asked the resource/queue manager?
> HTH
> Giuseppe
>
> Quoting Joe Page <J.Page3 at lboro.ac.uk>:
>
>> Hi,
>>
>> I am sending this message as I am running into some issues when
>> trying to perform a pw.x calculation.
>>
>> I am using a hybrid approach, with the flag input_dft = b3lpyv1r.
>> The first SCF calculation can be performed, however when it comes to
>> introducing the hybrid correction the program stops with the mpi
>> exit code 9 error.
>>
>> I am running this on a HPC, and I am using 100+ processors. I am
>> also using the pw.x -pd .true flag in the command line.
>>
>> I normally up the number of processors needed when I encounter this
>> issue, however I have a similar calculations, similar structure but
>> same hybrid correction, that uses a lot less processors to run.
>>
>> Is there something I am missing?
>>
>> Thank you for your help, and if there is anything else I can provide
>> that will aid in this request I will try to provide it.
>>
>> Thanks you again
>>
>> Joseph Page
>> Research Associate
>> Department of Physics
>> Loughborough University
>> j.page3 at lboro.ac.uk
>
>
>
> GIUSEPPE MATTIOLI
> CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
> Via Salaria Km 29,300 - C.P. 10
> I-00015 - Monterotondo Scalo (RM)
> Mob (*preferred*) +39 373 7305625
> Tel + 39 06 90672342 - Fax +39 06 90672316
> E-mail: <giuseppe.mattioli at ism.cnr.it>
>
> _______________________________________________
> 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<http://www.max-centre.eu>)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users



GIUSEPPE MATTIOLI
CNR - ISTITUTO DI STRUTTURA DELLA MATERIA
Via Salaria Km 29,300 - C.P. 10
I-00015 - Monterotondo Scalo (RM)
Mob (*preferred*) +39 373 7305625
Tel + 39 06 90672342 - Fax +39 06 90672316
E-mail: <giuseppe.mattioli at ism.cnr.it>