[Pw_forum] timing vs QE version vs pools
Giovanni Cantele
Giovanni.Cantele at na.infn.it
Mon Mar 9 14:33:59 CET 2009
Marino Vetuschi Zuccolini wrote:
> Dear Giovanni,
> I'm testing our system with more or less the same configuration. Can
> you send me the input you used in testing your cluster?
>
> Many thanks.
>
> m.
>
warning: k point list just for testing purposes (I just tried to get
those specific numbers
of k-points)
input:
&CONTROL
calculation =
'scf'
title =
'graphene'
restart_mode =
'from_scratch'
outdir =
'/scratch/Giovanni/bench_graph4x4'
prefix =
'C'
pseudo_dir =
'/home/nm_settings/software/CODES/Quantum-ESPRESSO/pseudo/'
/
&SYSTEM
ibrav =
4
celldm(1) =
18.6
celldm(3) =
1.36250
nat =
32
ntyp =
1
ecutwfc =
30.D0
ecutrho =
180.D0
occupations =
'smearing'
smearing =
'mv'
degauss =
0.007
nosym =
.true.
/
&ELECTRONS
/
ATOMIC_SPECIES
C 12.0107
C.pbe-rrkjus.UPF
ATOMIC_POSITIONS { Angstrom
}
C 0.000000000 0.000000000
1.676333520
C 1.230336530 0.710335120
1.676333520
C 18.455047875 2.131005372
1.676333520
C 0.000000005 2.841340492
1.676333520
C 17.224711350 4.262010743
1.676333520
C 18.455047880 4.972345863
1.676333520
C 15.994374825 6.393016115
1.676333520
C 17.224711355 7.103351235
1.676333520
C 2.460673050 0.000000000
1.676333520
C 3.691009580 0.710335120
1.676333520
C 1.230336525 2.131005372
1.676333520
C 2.460673055 2.841340492
1.676333520
C 0.000000000 4.262010743
1.676333520
C 1.230336530 4.972345863
1.676333520
C 18.455047875 6.393016115
1.676333520
C 0.000000005 7.103351235
1.676333520
C 4.921346100 0.000000000
1.676333520
C 6.151682630 0.710335120
1.676333520
C 3.691009575 2.131005372
1.676333520
C 4.921346105 2.841340492
1.676333520
C 2.460673050 4.262010743
1.676333520
C 3.691009580 4.972345863
1.676333520
C 1.230336525 6.393016115
1.676333520
C 2.460673055 7.103351235
1.676333520
C 7.382019150 0.000000000
1.676333520
C 8.612355680 0.710335120
1.676333520
C 6.151682625 2.131005372
1.676333520
C 7.382019155 2.841340492
1.676333520
C 4.921346100 4.262010743
1.676333520
C 6.151682630 4.972345863
1.676333520
C 3.691009575 6.393016115
1.676333520
C 4.921346105 7.103351235 1.676333520
The k-point list follows, these are the two tests (32 kpts and 64 kpts)
I did:
K_POINTS { crystal
}
32
0.0000000 0.0000000 0.0000000
1.00
0.2500000 0.1443376 0.0000000
1.00
0.5000000 0.2886751 0.0000000
1.00
0.7500000 0.4330127 0.0000000
1.00
0.0000000 0.2886751 0.0000000
1.00
0.2500000 0.4330127 0.0000000
1.00
0.5000000 0.5773503 0.0000000
1.00
0.7500000 0.7216878 0.0000000 1.00
0.0000000 0.5773503 0.0000000 1.00
0.2500000 0.7216878 0.0000000 1.00
0.5000000 0.8660254 0.0000000 1.00
0.7500000 1.0103630 0.0000000 1.00
0.0000000 0.8660254 0.0000000 1.00
0.2500000 1.0103630 0.0000000 1.00
0.5000000 1.1547005 0.0000000 1.00
0.7500000 1.2990381 0.0000000 1.00
0.0000000 0.0000000 0.3669725 1.00
0.2500000 0.1443376 0.3669725 1.00
0.5000000 0.2886751 0.3669725 1.00
0.7500000 0.4330127 0.3669725 1.00
0.0000000 0.2886751 0.3669725 1.00
0.2500000 0.4330127 0.3669725 1.00
0.5000000 0.5773503 0.3669725 1.00
0.7500000 0.7216878 0.3669725 1.00
0.0000000 0.5773503 0.3669725 1.00
0.2500000 0.7216878 0.3669725 1.00
0.5000000 0.8660254 0.3669725 1.00
0.7500000 1.0103630 0.3669725 1.00
0.0000000 0.8660254 0.3669725 1.00
0.2500000 1.0103630 0.3669725 1.00
0.5000000 1.1547005 0.3669725 1.00
0.7500000 1.2990381 0.3669725 1.00
K_POINTS { crystal
}
64
0.0000000 0.0000000 0.0000000
1.00
0.1250000 0.0721688 0.0000000
1.00
0.2500000 0.1443376 0.0000000
1.00
0.3750000 0.2165064 0.0000000
1.00
0.5000000 0.2886751 0.0000000
1.00
0.6250000 0.3608439 0.0000000
1.00
0.7500000 0.4330127 0.0000000
1.00
0.8750000 0.5051815 0.0000000
1.00
0.0000000 0.1443376 0.0000000
1.00
0.1250000 0.2165064 0.0000000
1.00
0.2500000 0.2886751 0.0000000
1.00
0.3750000 0.3608439 0.0000000
1.00
0.5000000 0.4330127 0.0000000
1.00
0.6250000 0.5051815 0.0000000
1.00
0.7500000 0.5773503 0.0000000
1.00
0.8750000 0.6495191 0.0000000
1.00
0.0000000 0.2886751 0.0000000
1.00
0.1250000 0.3608439 0.0000000
1.00
0.2500000 0.4330127 0.0000000
1.00
0.3750000 0.5051815 0.0000000
1.00
0.5000000 0.5773503 0.0000000
1.00
0.6250000 0.6495191 0.0000000
1.00
0.7500000 0.7216878 0.0000000
1.00
0.8750000 0.7938566 0.0000000
1.00
0.0000000 0.4330127 0.0000000
1.00
0.1250000 0.5051815 0.0000000
1.00
0.2500000 0.5773503 0.0000000
1.00
0.3750000 0.6495191 0.0000000
1.00
0.5000000 0.7216878 0.0000000
1.00
0.6250000 0.7938566 0.0000000
1.00
0.7500000 0.8660254 0.0000000
1.00
0.8750000 0.9381942 0.0000000
1.00
0.0000000 0.5773503 0.0000000
1.00
0.1250000 0.6495191 0.0000000
1.00
0.2500000 0.7216878 0.0000000
1.00
0.3750000 0.7938566 0.0000000
1.00
0.5000000 0.8660254 0.0000000
1.00
0.6250000 0.9381942 0.0000000
1.00
0.7500000 1.0103630 0.0000000
1.00
0.8750000 1.0825318 0.0000000 1.00
0.0000000 0.7216878 0.0000000 1.00
0.1250000 0.7938566 0.0000000 1.00
0.2500000 0.8660254 0.0000000 1.00
0.3750000 0.9381942 0.0000000 1.00
0.5000000 1.0103630 0.0000000 1.00
0.6250000 1.0825318 0.0000000 1.00
0.7500000 1.1547005 0.0000000 1.00
0.8750000 1.2268693 0.0000000 1.00
0.0000000 0.8660254 0.0000000 1.00
0.1250000 0.9381942 0.0000000 1.00
0.2500000 1.0103630 0.0000000 1.00
0.3750000 1.0825318 0.0000000 1.00
0.5000000 1.1547005 0.0000000 1.00
0.6250000 1.2268693 0.0000000 1.00
0.7500000 1.2990381 0.0000000 1.00
0.8750000 1.3712069 0.0000000 1.00
0.0000000 1.0103630 0.0000000 1.00
0.1250000 1.0825318 0.0000000 1.00
0.2500000 1.1547005 0.0000000 1.00
0.3750000 1.2268693 0.0000000 1.00
0.5000000 1.2990381 0.0000000 1.00
0.6250000 1.3712069 0.0000000 1.00
0.7500000 1.4433757 0.0000000 1.00
0.8750000 1.5155445 0.0000000 1.00
Please let me know your the results of your tests, I would be curious to
compare with mines.
Giovanni
> On 6 Mar 2009, at 18:15, Giovanni Cantele wrote:
>
>
>> Dear all,
>>
>> I just finished some timing tests relative to QE, stimulated by recent
>> discussions in this forum.
>> I would like to share these tests, and ask a couple of simple (?)
>> questions.
>>
>> Up to now, even if not supported by systematic tests, my feeling has
>> been that
>> for the kind of runs I'm doing on the kind of cluster(s) I'm using,
>> the
>> following rules hold:
>>
>> i) if using a relatively "small" number of CPUs, pools are rather
>> ineffective, I often realized that
>> using pools actually increased job execution time
>>
>> ii) diagonalization parallelism (-ndiag XX) is ineffective as well
>>
>> iii) version 3.2.3 is "better" than 4.0.4
>>
>>
>
> ...
>
> *******************************************************
> Marino Vetuschi Zuccolini
> zucco at dipteris.unige.it
> Researcher / Geochemist
> Laboratory of Geochemistry
>
> DIPartimento per lo studio della TErra e delle sue RISorse -
> Università di Genova
> Tel. ++39 010 3538136 Fax. ++39 010 352169
> Corso Europa 26, 16132 - Genova - Italy
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum
>
>
--
Dr. Giovanni Cantele
Coherentia CNR-INFM and Dipartimento di Scienze Fisiche
Universita' di Napoli "Federico II"
Complesso Universitario di Monte S. Angelo - Ed. 6
Via Cintia, I-80126, Napoli, Italy
Phone: +39 081 676910
Fax: +39 081 676346
E-mail: giovanni.cantele at cnr.it
giovanni.cantele at na.infn.it
Web: http://people.na.infn.it/~cantele
Research Group: http://www.nanomat.unina.it
More information about the Pw_forum
mailing list