My Community

Dear administer:
I compiled parallel edition of vasp4.6 using mpif90 without problem although it showed me some warning message. When running it, it aborted very soon and showed me error message as:

forrtl: severe (71): integer divide by zero

I really have no idea on what to do now. Anyone have similar experience and can help? Thanks a lot.

sorry, I don't have any idea either. This information is not enough to tell anything.
did the same job run serial,
where did vasp crash
what were the compiler warnings

I could not compile the serial edition of it when using ifort. It showed me following error message:
-L/usr/local/mkl-7.0.17/mkl70/lib/32 -lmkl -lmkl_lapack32 -lmkl_lapack64 -lguide -lmkl_p4 -lpthread
mpi.o(.text+0x24): In function `m_init':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:204: undefined reference to `mpi_init_'
mpi.o(.text+0xc1):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:216: undefined reference to `mpi_comm_rank_'
mpi.o(.text+0x118):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:221: undefined reference to `mpi_comm_size_'
mpi.o(.text+0x2c9): In function `m_divide':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:257: undefined reference to `mpi_cart_create_'
mpi.o(.text+0x32d):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:265: undefined reference to `mpi_cart_sub_'
mpi.o(.text+0x391):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:273: undefined reference to `mpi_cart_sub_'
mpi.o(.text+0x4a4): In function `m_initc':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:296: undefined reference to `mpi_comm_rank_'
mpi.o(.text+0x4ee):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:300: undefined reference to `mpi_comm_size_'
mpi.o(.text+0x545):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:307: undefined reference to `mpi_barrier_'
mpi.o(.text+0x616): In function `m_exit':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:357: undefined reference to `mpi_finalize_'
mpi.o(.text+0x6f2): In function `m_stop':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:383: undefined reference to `mpi_abort_'
mpi.o(.text+0x7d4): In function `m_stop_ierr':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:401: undefined reference to `mpi_abort_'
mpi.o(.text+0x874): In function `m_send_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:430: undefined reference to `mpi_send_'
mpi.o(.text+0x918): In function `m_recv_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:455: undefined reference to `mpi_recv_'
mpi.o(.text+0xad7): In function `m_sum_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:502: undefined reference to `mpi_allreduce_'
mpi.o(.text+0xceb): In function `m_max_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:548: undefined reference to `mpi_allreduce_'
mpi.o(.text+0xf0d): In function `m_max_d':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:597: undefined reference to `mpi_allreduce_'
mpi.o(.text+0x1135): In function `m_sumb_d':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:648: undefined reference to `mpi_allreduce_'
mpi.o(.text+0x1577): In function `m_sumb_z':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:734: undefined reference to `mpi_allreduce_'
mpi.o(.text+0x1704): In function `m_bcast_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:780: undefined reference to `mpi_bcast_'
mpi.o(.text+0x1837): In function `m_bcast_i_from':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:815: undefined reference to `mpi_bcast_'
mpi.o(.text+0x196a): In function `m_bcast_d':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:849: undefined reference to `mpi_bcast_'
mpi.o(.text+0x1a9e): In function `m_bcast_z':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:884: undefined reference to `mpi_bcast_'
mpi.o(.text+0x1bd1): In function `m_bcast_z_from':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:920: undefined reference to `mpi_bcast_'
mpi.o(.text+0x1fb2): In function `m_alltoallv_z':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:992: undefined reference to `mpi_irecv_'
mpi.o(.text+0x20da):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1009: undefined reference to `mpi_isend_'
mpi.o(.text+0x21f7):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1027: undefined reference to `mpi_waitall_'
mpi.o(.text+0x2366): In function `m_alltoall_i':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1069: undefined reference to `mpi_alltoallv_'
mpi.o(.text+0x242a): In function `m_alltoallv_raddr':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1100: undefined reference to `mpi_alltoall_'
mpi.o(.text+0x2526): In function `m_alltoallv_simple':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1136: undefined reference to `mpi_alltoall_'
mpi.o(.text+0x27cb): In function `m_alltoall_d':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1222: undefined reference to `mpi_irecv_'
mpi.o(.text+0x28ac):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1233: undefined reference to `mpi_isend_'
mpi.o(.text+0x297a):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1245: undefined reference to `mpi_waitall_'
mpi.o(.text+0x2bf4): In function `m_alltoall_d_async':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1332: undefined reference to `mpi_isend_'
mpi.o(.text+0x2c7d):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1336: undefined reference to `mpi_irecv_'
mpi.o(.text+0x2d7c): In function `m_alltoall_wait':
/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1359: undefined reference to `mpi_waitall_'
mpi.o(.text+0x2dd4):/e/cwr0316/vasp/src/vasp.4.6/mpi.f:1362: undefined reference to `mpi_waitall_'
paw.o(.text+0x52ce): In function `rd_rho_paw':
/e/cwr0316/vasp/src/vasp.4.6/paw.f:702: undefined reference to `mpi_barrier_'
electron.o(.text+0xa3d): In function `elmin':
/e/cwr0316/vasp/src/vasp.4.6/electron.f:286: undefined reference to `mpi_barrier_'
fft3dlib.o(.text+0x52b1): In function `ftrigc_':
: undefined reference to `vmldSinCos2'
fft3dlib.o(.text+0x541d): In function `ftrigr_':
: undefined reference to `vmldSinCos2'
make: *** [vasp] Error 1

I could not understand its meaning. So I decided to compile a parallel edition.

The parallel edition could be compiled and its warnings is about fft3dlib.f90, showing that some loops could not be vectorized like following:
fft3dlib.f90(4361) : (col. 9) remark: loop was not vectorized: existence of vector dependence.
fft3dlib.f90(4261) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4282) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4205) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4224) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4233) : (col. 9) remark: vector dependence: assumed FLOW dependence between b line 4235 and b line 4235.
fft3dlib.f90(4233) : (col. 9) remark: vector dependence: assumed FLOW dependence between a line 4234 and a line 4234.
fft3dlib.f90(4233) : (col. 9) remark: loop was not vectorized: existence of vector dependence.
fft3dlib.f90(4159) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4173) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4180) : (col. 6) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4725) : (col. 9) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4737) : (col. 9) remark: loop was not vectorized: not inner loop.
fft3dlib.f90(4752) : (col. 12) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4773) : (col. 12) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4818) : (col. 9) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4828) : (col. 9) remark: loop was not vectorized: not inner loop.
fft3dlib.f90(4843) : (col. 12) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4864) : (col. 12) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4895) : (col. 9) remark: loop was not vectorized: not inner loop.
fft3dlib.f90(4898) : (col. 12) remark: loop was not vectorized: subscript too complex.
fft3dlib.f90(4919) : (col. 9) remark: loop was not vectorized: not inner loop.
fft3dlib.f90(4922) : (col. 25) remark: loop was not vectorized: unsupported loop structure.

but it finished compiling. When running it, it crashed after showing the following content in the outcar file:

The dynamic configuration has the point symmetry C_1 .


KPOINTS: single kpoint
k-points in cartesian coordinates





using vasp > log.04 command, I got following log.04 results:

vasp.4.6.31 08Feb07 complex
POSCAR found : 2 types and 96 ions
LDA part: xc-table for Ceperly-Alder, standard interpolation

-----------------------------------------------------------------------------
| |
| W W AA RRRRR N N II N N GGGG !!! |
| W W A A R R NN N II NN N G G !!! |
| W W A A R R N N N II N N N G !!! |
| W WW W AAAAAA RRRRR N N N II N N N G GGG ! |
| WW WW A A R R N NN II N NN G G |
| W W A A R R N N II N N GGGG !!! |
| |
| VASP found 285 degrees of freedom |
| the temperature will equal 2*E(kin)/ (degrees of freedom) |
| this differs from previous releases, where T was 2*E(kin)/(3 NIONS). |
| The new definition is more consistent |
| |
-----------------------------------------------------------------------------

while the error message from the simulation is like following:
forrtl: severe (71): integer divide by zero
Image PC Routine Line Source
vaspmpi 080C38F9 Unknown Unknown Unknown
vaspmpi 080C2EA0 Unknown Unknown Unknown
vaspmpi 08056D52 Unknown Unknown Unknown
vaspmpi 0804C208 Unknown Unknown Unknown
libc.so.6 40B99BBA Unknown Unknown Unknown
vaspmpi 0804C0C1 Unknown Unknown Unknown
cp: cannot stat `TMPCAR': No such file or directory


It is so strange. I hope those information could help. Thanks a lot.

1) the fft- lib related messages are only parallelization remarks.
2) you seem to have a severe problem with the parallelization software you have installed. please contact your sys admit to get some information about that. The part of the makefile which generates parallel executables usually runs straightforwardly. All you have to do is set the $MPI shell variable correctly (in your .$SHELL-rc file) and use the appropriate precompiler options (in your makefile) and parallelization tools (on you computer system).
3) vmldSinCos2 is a routine of the libsvml (32-bit version) please check where it is installed and include it you your BLAS links