59 lines
1.6 KiB
Plaintext
59 lines
1.6 KiB
Plaintext
"LAMMPS WWW Site"_lws - "LAMMPS Documentation"_ld - "LAMMPS Commands"_lc :c
|
|
|
|
:link(lws,http://lammps.sandia.gov)
|
|
:link(ld,Manual.html)
|
|
:link(lc,Section_commands.html#comm)
|
|
|
|
:line
|
|
|
|
nthreads command :h3
|
|
|
|
[Syntax:]
|
|
|
|
nthreads n :pre
|
|
|
|
n = # of threads to be used per MPI task :ul
|
|
|
|
[Examples:]
|
|
|
|
nthreads 1
|
|
nthreads 6
|
|
nthreads * :pre
|
|
|
|
[Description:]
|
|
|
|
Specify how many threads are to be used per MPI task in a hybrid MPI/multi-threading environment.processors are mapped as a 3d logical grid to the global
|
|
simulation box.
|
|
|
|
When this command has not been specified, LAMMPS will choose n according
|
|
to the defaults of the threading environment, e.g. in OpenMP according to
|
|
the OMP_NUM_THREADS environment variable or using all available processors
|
|
if that is not set.
|
|
|
|
For a package like LAMMPS, threading is typically not as efficient as
|
|
a distributed data parallelization via domain decomposition using MPI
|
|
message passing. Thus this keyword is mostly of use when MPI parallelization
|
|
is becoming inefficient, e.g. when using a "slow" network (like GigE with
|
|
TCP/IP) or when reaching the limit of scaling.
|
|
|
|
[Restrictions:]
|
|
|
|
This command cannot be used after the simulation box is defined by a
|
|
"read_data"_read_data.html or "create_box"_create_box.html command.
|
|
|
|
This command only has an effect and is only of use if LAMMPS was
|
|
built with the "user-openmp" package included. See the "Making
|
|
LAMMPS"_Section_start.html#2_3 section for more info.
|
|
|
|
[Related commands:]
|
|
|
|
"processors"_processors.html
|
|
|
|
[Default:]
|
|
|
|
nthreads * :pre
|
|
|
|
or in case multi-threading support is not compiled in:
|
|
|
|
nthreads 1 :pre
|