[DFTB-Plus-User] Problem of parallel computing with openmpi
yokoi-mp.pse
yokoi at mp.pse.nagoya-u.ac.jp
Thu May 11 15:35:10 CEST 2017
Dear DFTB+ developers,
I am a new user of DFTB+, and now I have a problem of parallel computing
with openmpi.
When I perform a MD calculation on 4 cores in one node, computational
time effectively decreases compared to a single-core calculation.
However, when the same MD calculation is done on 8 cores of two nodes,
its computational time become 2-3 times longer than the case of 4 cores
in one node.
I think that the calculated simulation box contains relatively a large
number of atoms (496 Si atoms) and the number of nodes is just two, and
thus communication overhead is not a cause. Would you give me advice
about possible causes of this problem?
My architecture and package version are
CPU: intel core i7
OS: CentOS 6.8
intel compiler and mkl version: 14.0.2
openmpi version: 1.6.5 and 1.8.8
DFTB+: dftb+.mpi-r4473
An input file of the MD simulation is also attached as follows.
Geometry = GenFormat {
<<< "config.dftb.in" # This simulation box contains 496 Si
atoms.
}
Driver = VelocityVerlet {
Steps = 100000
TimeStep [Femtosecond] = 1.0
Thermostat = Andersen {
Temperature [Kelvin] = 1000.0
ReselectProbability = 0.200000000000000
ReselectIndividually = No
AdaptFillingTemp = Yes
}
OutputPrefix = "geo_end"
}
Hamiltonian = DFTB {
SCC = No
MaxAngularMomentum = {
Si = "p"
}
SlaterKosterFiles = {
Si-Si = "./Si-Si.skf"
}
KPointsAndWeights = SupercellFolding {
1 0 0
0 4 0
0 0 6
0.0 0.0 0.0
}
}
Options = {
WriteAutotestTag = Yes
AtomResolvedEnergies = Yes
}
ParserOptions = {
ParserVersion = 3
}
Sincerely,
tatsu
More information about the DFTB-Plus-User
mailing list