array out of bound error

Questions regarding the compilation of VASP on various platforms: hardware, compilers and libraries, etc.

Moderators: Global Moderator, Moderator

Locked
Message
Author
sickboy
Newbie
Newbie
Posts: 1
Joined: Wed May 15, 2013 9:04 am

array out of bound error

#1 Post by sickboy » Wed May 15, 2013 9:31 am

Dear all,
I have met a segment fault when I was tring to run vasp 5.3.3. The binary file was compiled with intel cluster studio 2012 (ifort 12.1.0 IMPI 4.0 update 3 MKL 10.3 Update 6). The following flags was used when I was compiling vasp
OFLAG=-O0
FFLAGS=-FR -names lowercase -assume byterecl
BLAS=-lmkl_blas95_lp64
LAPACK=-lmkl_lapack95_lp64
BLACS=-lmkl_blacs_intelmpi_lp64 -lpthread -lm
SCA=-lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core $(BLACS)
LIB=-L../vasp.5.lib -ldmy ../vasp.5.lib/linpack_double.o $(BLAS) $(LAPACK) $(SCA)
CPP=$(CPP_) -DMPI -DHOST=\"LinuxIFC\" -DIFC -DCACHE_SIZE=4000 -DPGF90 -Davoidalloc -DNGZhalf -DMPI_BLOCK=8000 -Duse_collective -DscaLAPACK
The other rest of makefile was copied from makefile.linux_p4_ifc
The compilation of vasp completed normally.
However, when I tries to run it in a specific job, it fails with such errors


forrtl: severe (408): fort: (2): Subscript #1 of the array TBLI has value 1 which is greater than the upper bound of 0

Image PC Routine Line Source
vasp5.3 000000000382AFCA Unknown Unknown Unknown
vasp5.3 0000000003829AC6 Unknown Unknown Unknown
vasp5.3 00000000037E83F0 Unknown Unknown Unknown
vasp5.3 000000000379945E Unknown Unknown Unknown
vasp5.3 0000000003799981 Unknown Unknown Unknown
vasp5.3 0000000003717772 mapset_ 271 fftmpi_map.F
vasp5.3 000000000041DEDF Unknown Unknown Unknown
vasp5.3 000000000040654C Unknown Unknown Unknown
libc.so.6 0000003D2401ECDD Unknown Unknown Unknown
vasp5.3 0000000000406449 Unknown Unknown Unknown


Apperantly, there is something wrong...
the INCAR file is pasted here
SYSTEM = O2
ISYM = 0
ISTART = 1 ! 0 = new job, 1 = restart
ISPIN = 2 ! spin polarisation 1 = no 2 = yes
PREC = normal ! [Low/Medium/High/Accurate]
ENCUT = 400 ! energy cutoff
EDIFF = 1.0E-5 ! [Global break cond. for the elec. SC-loop] Energy
EDIFFG = -0.1 ! convergence criterion for forces
NELM = 400 ! max iteraci v SC loop
NELMIN = 3 ! min # iterations in SC loop
!ISIF = 0 ! [Stress tensor, 0 = not, 3 = relax latt+ions] p. 60
NWRITE = 1 ! how much is written out
LCHARG = .True.! write restart charge file
LWAVE = .True. ! write restart wf file
ISMEAR = 0; SIGMA = 0.05 #-5 for accurate total energy and DOS 0 for large cell 1 for metal
NSW = 0 ! max # steps
LREAL = A ! aproximace, T zpomaluje, ale presnejsi
IBRION= 2 ! 0 MD 1 Brodyn (opt,TS) 2 CG (opt) 3 damped MD(opt) 5 FC
POTIM=0.2 ! ion step size/ time step (fs) for MD

LASPH = .TRUE. ! non-spherical contributions. neccessary for vdw
NPAR = 2 ! adjust to number of processors; best on berni up to 8 CPUs
!IALGO = 38 ! RMM-DIIS algorithm for electrons (Davidson: 38)
!ALGO=Fast

#PBE0
#LHFCALC = .TRUE. ; ALGO = Damped ; TIME = 0.5
#AEXX=0.25

#HSE06
PRECFOCK= A
LHFCALC = .TRUE. ; ALGO = Damped ; TIME = 0.5
HFSCREEN = 0.3

I has also tested my vasp5.3 binary on another job which could be normally completed by vasp5.2, it also failed with same errors.
Could any one help me out?

<span class='smallblacktext'>[ Edited ]</span>
Last edited by sickboy on Wed May 15, 2013 9:31 am, edited 1 time in total.

Locked