|
Hello,
Thank you very much for your kind reply and
comments~!
Best regards,
Yunhee Chang
----- Original Message -----
Sent: Thursday, May 18, 2006 3:48 PM
Subject: Re: [abinit-forum] Problems in the
IBM p690 machine using abinit-4.6.5
Hi,
This may not be related yet if other programs have also
some problem to allocate wide array of memory ...
Three years ago IBM
provided an AIX machine to the lab where I was working (power-4, 8GB ram). It
happens that some parameters of AIX were not really well chosen. For instance
it was not possible for an user to have more than 128 processes running on the
machine and for a process to allocate more than 256 MB of memory. So if for
instance a simple c programm doing almost only a big malloc faile in the same
way, you might wan't to tune the installed AIX because this kind of parameters
(memlimits, proclimits...) are purely arbitrary.
Otherwise you may
alsobe able to split a little bit more the memory consumption of your
calculation using more processes.
regards
PMA
On 5/18/06, yhjang@pusan.ac.kr <yhjang@pusan.ac.kr> wrote:
Dear
Abinit users,
With the 4.6.5 abinip version on our supercomputer
of IBM p690 SP4 machines.
OS is AIX and fortran compiler is
mxlf90.
When the reading an input file and check the memory needs,
I
get an error during the defining the memory need of the present
run.
Before on our cluster of Intel P4 machine running Linux, there
was no problems.
Has anyone else had
this?
=======================================================================
Biggest
array : cg(disk), with 571.3790 MBytes.
-P-0000
-P-0000 memana : ERROR
-
-P-0000 Test failed to allocate an array of 571.379 Mbytes
-P-0000
It is not worth to continue
-P-0000 Action : modify input variable to fit
the available memory,
-P-0000 or increase limit on maximal array
size.
-P-0000
-P-0000 leave_new : decision taken to exit ...
-P-0000 leave_new : synchronization done...
-P-0000 leave_new :
exiting...
=========================================================================
I
attatch the log file
below~!
==========================================================================
Version 4.6.5 of ABINIT
(MPI version, prepared for a ibm
computer)
Copyright (C) 1998-2005 ABINIT group .
ABINIT comes with
ABSOLUTELY NO WARRANTY.
It is free software, and you are welcome to
redistribute it
under certain conditions (GNU General Public
License,
see ~ABINIT/Infos/copyright or http://www.gnu.org/copyleft/gpl.txt).
ABINIT
is a project of the Universite Catholique de Louvain,
ABINIT is a
project of the Universite Catholique de Louvain,
Corning Inc. and other
collaborators, see ~ABINIT/Infos/contributors.
Please read
~ABINIT/Infos/acknowledgments.html for suggested
acknowledgments of the
ABINIT effort.
For more information, see http://www.abinit.org .
Starting
date : Thu 18 May 2006.
- input file -> a4x1L.in
- output file
-> a4x1L.outB
- root for input files -> a4x1Li
- root for
output files -> a4x1Lo
instrng : 184 lines of input have been
read
iofn2 : Please give name of formatted atomic psp file
iofn2 :
for atom type 1 , psp file is /inst2/x147khc/ychang/abinit-
4.6.5/Psps/GGA_FHI/49-In.GGA.fhi
read the values zionpsp= 3.0 , pspcod= 6
, lmax= 3
iofn2 : Please give name of formatted atomic psp
file
iofn2 : for atom type 2 , psp file is
/inst2/x147khc/ychang/abinit-4.6.5 /Psps/GGA_FHI/14-Si.GGA.fhi
read the
values zionpsp= 4.0 , pspcod= 6 , lmax= 3
iofn2 : Please give name of
formatted atomic psp file
iofn2 : for atom type 3 , psp file is
/inst2/x147khc/ychang/abinit-4.6.5/Psps/GGA_FHI/14- Si.GGA.fhi
read the
values zionpsp= 4.0 , pspcod= 6 , lmax= 3
iofn2 : Please give name of
formatted atomic psp file
iofn2 : for atom type 4 , psp file is
/inst2/x147khc/ychang/abinit-4.6.5/Psps/GGA_FHI/01-H.GGA.fhi
read the
values zionpsp= 1.0 , pspcod= 6 , lmax= 3
iofn2 : deduce mpsang = 4,
n1xccc =2501.
-P-0000 leave_test : synchronization
done...
invars1m : enter jdtset= 0
ingeo : takes atomic
coordinates from input array xangst
invars1m : enter jdtset= 0
ingeo
: takes atomic coordinates from input array xangst
symanal : COMMENT
-
The Bravais lattice determined only from the primitive
vectors,
bravais(1)= 3, is more symmetric
than the real one, iholohedry= 1,
obtained by taking into
account the atomic positions.
symspgr : the
symmetry operation no. 1 is the identity
symspgr : spgroup= 1 P1
(=C1^1)
inkpts: Sum of 16 k point weights is 16.000000
With present
occopt= 1 , renormalize it to one
inkpts : istwfk preprocessed, gives
following first values (max. 6): 1 1 1 1 1 1
invars1: mkmem undefined in
the input file. Use default mkmem = nkpt
invars1: With nkpt_me= 8 and
mkmem = 16, ground state wf handled in core.
Resetting mkmem to nkpt_me
to save memory space.
invars1: mkqmem undefined in the input file. Use
default mkqmem = nkpt
invars1: With nkpt_me= 8 and mkqmem = 16, ground
state wf handled in core.
Resetting mkqmem to nkpt_me to save memory
space.
invars1: mk1mem undefined in the input file. Use default mk1mem =
nkpt
invars1: With nkpt_me= 8 and mk1mem = 16, ground state wf handled in
core.
Resetting mk1mem to nkpt_me to save memory space.
Symmetries : space group P1 (# 1); Bravais aP (primitive
triclinic)
invars2: take the default value of fband=
1.25000000E-01
invars2: COMMENT -
When ionmov==4 and iscf==5 or 6,
iprcch must be 3.
Set iprcch to 3.
inkpts: Sum of 16 k point weights
is 16.000000
With present occopt= 1 , renormalize it to one
inkpts
: istwfk preprocessed, gives following first values (max. 6): 1 1 1 1 1
1
chkneu : initialized the occupation numbers for occopt= 1
spin-unpolarized case :
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00
spin-unpolarized case :
2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 2.00 2.00 2.00
2.00 2.00 2.00 2.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00
0.00 0.00 0.00 0.00 0.00
For input ecut=
8.000000E+00 best grid ngfft= 80 40 160
max ecut= 8.903865E+00
getng:
value of mgfft= 160 and nfft= 512000
getng: values of
ngfft(4),ngfft(5),ngfft(6) 81 41 160
getmpw: optimal value of mpw=
23760
iofn2 : deduce lmnmax = 16, lnmax = 4,
lmnmaxso= 16,
lnmaxso= 4.
memory : analysis of memory
needs
================================================================================
Values of the parameters that define the memory need of the present
run
intxc = 0 ionmov = 4 iscf = 5 ixc = 11
lmnmax = 4 lnmax = 4 mband
= 197 mffmem = 1
P mgfft = 160 mkmem = 8 mpssoang= 4 mpw =
23760
mqgrid = 1201 natom = 100 nfft = 512000 nkpt = 16
nloalg = 4
nspden = 1 nspinor = 1 nsppol = 1
nsym = 1 n1xccc = 2501 ntypat = 4
occopt =
1
================================================================================
P
This job should need less than 705.942 Mbytes of
memory.
================================================================================
P
This job should need less than 705.942 Mbytes of memory.
Rough estimation
(10% accuracy) of disk space for files :
WF disk file : 1142.756 Mbytes
; DEN or POT disk file : 3.908
Mbytes.
================================================================================
Biggest
array : cg(disk), with 571.3790 MBytes.
-P-0000
-P-0000 memana :
ERROR -
-P-0000 Test failed to allocate an array of 571.379
Mbytes
-P-0000 It is not worth to continue
-P-0000 Action : modify
input variable to fit the available memory,
-P-0000 or increase limit on
maximal array size.
-P-0000
-P-0000 leave_new : decision taken to
exit ...
-P-0000 leave_new : synchronization done...
-P-0000 leave_new
:
exiting...
=============================================================================
Thanks for help,
Yunhee Chang
--
Pierre-Matthieu Anglade
|
