[QE-users] (no subject)

문태환 th4512 at cau.ac.kr
Fri Sep 7 05:43:50 CEST 2018


Dear Minho Lee,

As you mentioned, we looked at the quota of 'swap' memory through the 'file' command.

The SWAP memory was allocated at 32 GB, but the actual memory was only 3 GB.

While checking the SCF's log file, I found something different from the previous one.

Processes (MPI) and Threads (openMP) used in the calculation are the same as 8 and 4 respectively,

Successful calculations before reinstalling the computer include the following:

     Parallel version (MPI & OpenMP), running on 32 processor cores
     Number of MPI processes: 8
     Threads / MPI process: 4
     R & G space division: proc / nbgrp / npool / nimage = 32
     Reading input from espresso.in

However, in the current log file, there are four nodes separated as follows.

     Parallel version (MPI & OpenMP), running on 4 processor cores
     Number of MPI processes: 1
     Threads / MPI process: 4

     MPI processes distributed on 1 nodes

This seems to be the cause of exceeding the memory of my computer .. Do you know why this happens?

Thank you!



-------------------------------------

Moon TaeHwan

Chung-Ang University

th4512 at cau.ac.kr

-------------------------------------


________________________________
보낸 사람: Minho Lee <mino at simulation.re.kr> 대신 users <users-bounces at lists.quantum-espresso.org>
보낸 날짜: 2018년 9월 7일 금요일 오전 10:29:00
받는 사람: Quantum Espresso users Forum
제목: Re: [QE-users] (no subject)

Dear TaeHwan Moon,

I’m sorry for the typo(ion_positions->atomic_positions )

Did you check your output?

In the standard output, you can find “estimated …..RAM ….” lines.

and you can check real memory usage of your computer via linux command “free”

If physical memory of your computer is not enough, “Swap” will be increased

Sincerly,

Minho Lee
------------------------
Minho Lee
CEO/Developer
T. 02-3293-0204 | C. 010-2770-4080
E. mino at virtuallab.co.kr | H. http://www.virtuallab.co.kr

On-demand Materials Simulation
https://www.materialssquare.com
On 7 Sep 2018, 9:51 AM +0900, 문태환 <th4512 at cau.ac.kr>, wrote:

Dear Minho Lee,

Thank you for your kind reply.

If you are talking about ATOMIC_POSITIONS, I thought I took a lot of text, so I excluded it when I wrote!

I have run the calculations by reducing the Processes (MPI) to 4, but the calculations are still going smoothly without exceeding the memory.

But my PC does not use CPU well and calculation speed is slow.

Like the Gaussian program,

Do you know how to control the memory while allocating a lot of CPU in the calculation?



-------------------------------------

Moon TaeHwan

Chung-Ang University

th4512 at cau.ac.kr

-------------------------------------


________________________________
보낸 사람: Minho Lee <mino at simulation.re.kr> 대신 users <users-bounces at lists.quantum-espresso.org>
보낸 날짜: 2018년 9월 6일 목요일 오후 8:17:16
받는 사람: users at lists.quantum-espresso.org; Quantum Espresso users Forum
제목: Re: [QE-users] (no subject)

Dear Taehwan Moon,

I can’t find the &ION namelist in your script,

I think you need to set ion_positions

Sincerely,

Minho Lee
------------------------
Minho Lee
CEO/Developer
T. 02-3293-0204 | C. 010-2770-4080
E. mino at virtuallab.co.kr | H. http://www.virtuallab.co.kr

On-demand Materials Simulation
https://www.materialssquare.com
On 2018년 9월 6일 PM 7:48 +0900, 문태환 <th4512 at cau.ac.kr>, wrote:

Hello! Quantum espresso users.
I am a beginner user of QE and Linux, and I am having a lot of computation difficulties.
I need to calculate the SCF, DOS, and Band structure using a CIF file with zero-dimensional organic-inorganic hybrid materials.
However, when the SCF calculation starts and after an hour or so has elapsed, the memory problem causes an error.
My computer has 32 core CPU and 64GB of memory.  In the case of the CPU, there is room for using 8 cores when executing calculations.
(Processes (MPI) and Threads (openMP) use 8 and 4, respectively.)

The input file looks like this:

Can you give me advice to solve this problem?
Please help me .. Thank you.


&CONTROL
    calculation  = "scf"
    max_seconds  =  8.64000e+04
    pseudo_dir   = "/home/user/.burai/.pseudopot"
    restart_mode = "from_scratch"
    tprnfor      = .FALSE.
    tstress      = .FALSE.
/

&SYSTEM
    a                         =  1.19962e+01
    b                         =  1.46765e+01
    c                         =  1.38899e+01
    constrained_magnetization = "none"
    cosac                     = -5.92053e-02
    degauss                   =  1.00000e-02
    ecutrho                   =  5.06000e+02
    ecutwfc                   =  5.62066e+01
    ibrav                     = -12
    lda_plus_u                = .FALSE.
    nat                       = 192
    nosym                     = .FALSE.
    nspin                     = 1
    ntyp                      = 5
    occupations               = "smearing"
    smearing                  = "guassian"
    starting_magnetization(1) =  0.00000e+00
/

&ELECTRONS
    conv_thr         =  1.00000e-06
    diago_david_ndim = 4
    diagonalization  = "david"
    electron_maxstep = 100
    mixing_beta      =  7.00000e-01
    mixing_mode      = "plain"
    mixing_ndim      = 8
    startingpot      = "atomic"
    startingwfc      = "atomic+random"
/

K_POINTS {automatic}
 3  3  3  0 0 0

ATOMIC_SPECIES
Bi    208.98038  Bi.pbe-dn-kjpaw_psl.1.0.0.UPF
Br     79.90400  Br.pbe-dn-kjpaw_psl.1.0.0.UPF
N      14.00674  N.pbe-n-kjpaw_psl.1.0.0.UPF
H       1.00794  H.pbe-kjpaw_psl.1.0.0.UPF
C      12.01070  C.pbe-n-kjpaw_psl.1.0.0.UPF




-------------------------------------

Moon TaeHwan

Chung-Ang University

th4512 at cau.ac.kr

-------------------------------------


_______________________________________________
users mailing list
users at lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
_______________________________________________
users mailing list
users at lists.quantum-espresso.org
https://lists.quantum-espresso.org/mailman/listinfo/users
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20180907/c9b43cc8/attachment.html>


More information about the users mailing list