Dear All,
I am calculating a slab of FASnI3 (972 atoms), 45 Ang thick plus 50 Ang of
vacuum with this siesta version
>Siesta Version : v4.1-b4
>Architecture : unknown
>Compiler version: GNU Fortran (GCC) 4.8.5 20150623 (Red Hat 4.8.5-36)
>Compiler flags : mpif90 -O2 -fPIC -ftree-vectorize
>PP flags : -DFC_HAVE_ABORT -DMPI -DSIESTA__DIAG_2STAGE
>Libraries : libsiestaLAPACK.a libsiestaBLAS.a
/opt/share/scalapack_gnu/libscalapack.a
>PARALLEL version
>* Running on 24 nodes in parallel
This is the initial .fdf I use and that, mutatis mutandis, works perfectly
for the bulk calculation (12 atoms):
SystemName surf010
SystemLabel surf010
NumberOfSpecies 5
NumberOfAtoms 972
XC.functional VDW
XC.authors VV
%block ChemicalSpeciesLabel
1 50 Sn
2 53 IGGAtm2
3 6 C
4 7 N
5 1 H
%endblock ChemicalSpeciesLabel
LatticeConstant 1.00 Ang
%block LatticeVectors
18.633 0. 0.
0. 94.5837 0.
0. 0. 25.8042
%endblock LatticeVectors
MeshCutoff 300 Ry
ElectronicTemperature 1500 K
UseSaveData .true.
MD.UseSaveXV .true.
DM.MixingWeight 0.02! 0.2
DM.NumberPulay 9 !3
DM.Tolerance 1.d-3 #1.d-5
SolutionMethod diagon
WriteDM .true.
WriteEigenvalues .true.
MaxSCFIterations 1000
PAO.BasisSize TZP
PAO.EnergyShift 0.01 Ry
MD.TypeOfRun CG
MD.VariableCell .false.
MD.RelaxCellOnly .false.
MD.MaxForceTol 0.04 ev/Ang ! 0.02 ev/Ang
MD.MaxStressTol 0.1 GPa
MD.NumCGsteps 3000
MD.FCDispl 0.03 Bohr
WriteCoorXmol .true.
WriteCoorMDXmol .true.
WriteMDXmol .true.
%block kgrid_Monkhorst_Pack
1 0 0 0.
0 1 0 0.
0 0 1 0.
%endblock kgrid_Monkhorst_Pack
AtomicCoordinatesFormat Ang
WriteCoorCerius .true.
%block AtomicCoordinatesAndAtomicSpecies
...
...
...
%endblock AtomicCoordinatesAndAtomicSpecies
And this the error I get immediately after launching the job
Setting up quadratic distribution...
ExtMesh (bp) on 0 = 154 x 136 x 151 = 3162544
PhiOnMesh: Number of (b)points on node 0 = 194568
PhiOnMesh: nlist on node 0 = 7689058
1637
rdiag: Error in Cholesky factorisation
Stopping Program from Node: 23
1637
rdiag: Error in Cholesky factorisation
Stopping Program from Node: 1
1637
rdiag: Error in Cholesky factorisation
Stopping Program from Node: 2
1637
rdiag: Error in Cholesky factorisation
Stopping Program from Node: 3
1637
rdiag: Error in Cholesky factorisation
Stopping Program from Node: 4
1637
rdiag: Error in Cholesky factorisation
...
...
Stopping Program from Node: 10
--------------------------------------------------------------------------
MPI_ABORT was invoked on rank 18 in communicator MPI COMMUNICATOR 3 CREATE
FROM 0
with errorcode 1.
NOTE: invoking MPI_ABORT causes Open MPI to kill all MPI processes.
You may or may not see output from other processes, depending on
exactly when Open MPI kills them.
--------------------------------------------------------------------------
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] PMIX ERROR: UNREACHABLE in file server/pmix_server.c at
line 2079
[node02:146174] 23 more processes have sent help message help-mpi-api.txt /
mpi-abort
[node02:146174] Set MCA parameter "orte_base_help_aggregate" to 0 to see
all help / error messages
I have tried all the possible solutions (Divide&Conquer,ParalleloverY).
The geometry is OK. Triple check. Everything worked perfectly for the bulk
case, as said...
My SysAdmin recommended me to contact you before proceeding since you
probably can better point out the origin of the issue.
Thanks in advance for your help/hints.
Best regards,
Giacomo Giorgi
P.S.: One further comment. I have noticed that in this version "UseSaveData
.true." is not sufficient per se to restart from previous XV file (It
only reads .DM file) and that "MD.UseSaveXV .true." must be specified
as well to make the geometry restart form previous XV file.
--
"Oltre le illusioni di Timbuctù e le gambe lunghe di Babalù c'era questa
strada...Questa strada zitta che vola via come una farfalla, una nostalgia,
nostalgia al gusto di curaçao...Forse un giorno meglio mi spiegherò"
(Paolo Conte, "Hemingway")
--
SIESTA is supported by the Spanish Research Agency (AEI) and by the European
H2020 MaX Centre of Excellence (http://www.max-centre.eu/)
