ABINIT parallelisation input variables:

List and description.


This document lists and provides the description of the name (keywords) of parallelisation input variables to be used in the main input file of the abinis code.

The new user is advised to read first the new user's guide, before reading the present file. It will be easier to discover the present file with the help of the tutorial.

When the user is sufficiently familiarized with ABINIT, the reading of the ~abinit/doc/users/tuning file might be useful. For response-function calculations using abinis, please read the response function help file

Copyright (C) 1998-2007 ABINIT group (DCA, XG, RC)
This file is distributed under the terms of the GNU General Public License, see ~abinit/COPYING or http://www.gnu.org/copyleft/gpl.txt .
For the initials of contributors, see ~abinit/doc/developers/contributors.txt .

Goto : ABINIT home Page | Suggested acknowledgments | List of input variables | Tutorial home page | Bibliography
Help files : New user's guide | Abinis (main) | Abinis (respfn) | Mrgddb | Anaddb | AIM (Bader) | Cut3D | Optic | Mrgscr
Files that describe other input variables:

Content of the file : alphabetical list of variables.


A.
B.
C.
D.
E.
F. fft_opt_lob  
G. gpara   gwpara  
H.
I.
J.
K.
L. localrdwf  
M.
N. ngroup_rf   npband   npfft  
O.
P. paral_rf  
Q.
R.
S.
T.
U.
V.
W.
X.
Y.
Z.




fft_opt_lob
Mnemonics: Fast Fourier Transform parallelisation - OPTion for LOB algorithm
Characteristic:
Variable type: integer
Default is 0.

Option for LOB algorithm, used in the FFT parallelisation, see npband, npfft.



Go to the top | Complete list of input variables
gpara
Mnemonics: G-vector parallelisation
Characteristic:
Variable type: integer
Default is 0.

(not active at present)



Go to the top | Complete list of input variables
gwpara
Mnemonics: GW parallelisation level
Characteristic:
Variable type: integer
Default is 1 (should become 2 later).

Option for GW parallelisation level, used in abinip. Check also localrdwf



Go to the top | Complete list of input variables
localrdwf
Mnemonics: LOCAL ReaD WaveFunctions
Characteristic: DEVELOP, PARALLEL
Variable type: integer
Default is 1.

This input variable is used only in abinip. If localrdwf=1, the input wavefunction disk file or KSS/SCR file in case of GW calculation, is read locally by each processor, while if localrdwf=0, only one processor reads it, and BCAST it to other processors.

The option localrdwf=0 is NOT allowed when mkmem==0 (or, for RF, when mkqmem==0, or mk1mem==0), that is, when the wavefunctions are stored on disk. This is still to be coded ...

In the case of a parallel computer with a unique file system, both options are as convenient for the user. However, if the I/O are slow compared to communications between processors, (e.g. for CRAY T3E machines), localrdwf=0 should be much more efficient; if you really need temporary disk storage, switch to localrdwf=1 ).

In the case of a cluster of nodes, with a different file system for each machine, the input wavefunction file must be available on all nodes if localrdwf=1, while it is needed only for the master node if localrdwf=0.



Go to the top | Complete list of input variables


ngroup_rf
Mnemonics: Define the number of groups for parallelization over perturbations
Characteristic: can even be specified separately for each dataset, parameter paral_rf is necessary
Variable type: integer
Default is 0.

This parameter is used in connection to the parallelization over perturbations. It defines the number of groups for distributing the perturbation-cases over the total number of available processors. The maximum number of groups is limited by the number of perturbation cases. The size of each group is again limited by the number k-points.



Go to the top | Complete list of input variables
npband
Mnemonics: Number of Processors for the BAND parallelisation
Characteristic:
Variable type: integer
Default is 1.

Relevant only for the band/FFT parallelisation. With this parallelisation, the work load is split over the two dimensions of a band/FFT grid. The different communications only occur along one dimension only.
npband gives the number of processors among which the work load over bands is shared. It is to be combined with npfft to give the total number of processors working on the band/FFT parallelisation. For nproc<=50, the best speed-up is achieved for npband=nproc and npfft=1 (which is not yet the defaut) For nproc<=50, the best speed-up is achieved for npband >= 4*npfft Flags of pre-compilation: This parallelisation is activated by the -DMPI_FFT flag. At present (v5.3), this parallelisation is not compatible with the k-point, spin or perturbation ones (the -DMPI flag). The keywords wfoptalg=4, nloalg=4, fftalg=401, iprcch=0, intxc=0, istwfk=1 and fft_opt_lob=2 have to be used for a band/FFpT parallelisation. Restriction: The Spin polarization as well as PAW method do not work. Allowed values: The values of the npband and npfft input variables have to be commensurate with the number of bands (nband) and the number of FFT planes along the 2nd and 3rd dimensions (ngfft(2),ngfft(3)), respectively. In addition, npfft should be smaller than min(ngfft(2),ngfft(3))/2.




Go to the top | Complete list of input variables
npfft
Mnemonics: Number of Processors for the FFT parallelisation
Characteristic:
Variable type: integer
Default is nproc.

Relevant only for the band/FFT parallelisation. With this parallelisation, the work load is split over the two dimensions of a band/FFT grid. The different communications only occur along one dimension only.
npfft gives the number of processors among which the work load over the FFT grid is shared. npfft and npband are combined to give the total number of processor (nproc) working on the band/FFT parallelisation. See npband for the additional information on the use of band/FFT parallelisation.
Note : npfft should be commensurate with ngfft(2) and ngfft(3)), and smaller than min(ngfft(2),ngfft(3))/2.




Go to the top | Complete list of input variables
paral_rf
Mnemonics: Activate parallelization over perturbations
Characteristic: can even be specified separately for each dataset, parameter ngroup_rf is necessary
Variable type: integer
Default is 0.

This parameter activates the parallelization over perturbations which can be used during RF-Calculation. It is possible to use this type of parallelization in combination to the parallelization over k-points.

Currently total energies calculated by groups, where the master process is not in, are saved in .status_LOGxxxx files.





Go to the top | Complete list of input variables
Goto : ABINIT home Page | Suggested acknowledgments | List of input variables | Tutorial home page | Bibliography
Help files : New user's guide | Abinis (main) | Abinis (respfn) | Mrgddb | Anaddb | AIM (Bader) | Cut3D | Optic | Mrgscr