<P>Note that if the "-sf omp" switch is used, it also issues a default
<AHREF ="package.html">package omp 0</A> command, which sets the number of threads
per MPI task via the OMP_NUM_THREADS environment variable.
</P>
<P>Using the "-pk" switch explicitly allows for direct setting of the
number of threads and additional options. Its syntax is the same as
the "package omp" command. See the <AHREF ="package.html">package</A> command doc
page for details, including the default values used for all its
options if it is not specified, and how to set the number of threads
via the OMP_NUM_THREADS environment variable if desired.
</P>
<P><B>Or run with the USER-OMP package by editing an input script:</B>
</P>
<P>The discussion above for the mpirun/mpiexec command, MPI tasks/node,
and threads/MPI task is the same.
</P>
<P>Use the <AHREF ="suffix.html">suffix omp</A> command, or you can explicitly add an
"omp" suffix to individual styles in your input script, e.g.
</P>
<PRE>pair_style lj/cut/omp 2.5
</PRE>
<P>You must also use the <AHREF ="package.html">package omp</A> command to enable the
USER-OMP package, unless the "-sf omp" or "-pk omp" <AHREF ="Section_start.html#start_7">command-line
switches</A> were used. It specifies how many
threads per MPI task to use, as well as other options. Its doc page
explains how to set the number of threads via an environment variable
if desired.
</P>
<P><B>Speed-ups to expect:</B>
</P>
<P>Depending on which styles are accelerated, you should look for a
reduction in the "Pair time", "Bond time", "KSpace time", and "Loop
time" values printed at the end of a run.
</P>
<P>You may see a small performance advantage (5 to 20%) when running a
USER-OMP style (in serial or parallel) with a single thread per MPI
task, versus running standard LAMMPS with its standard
(un-accelerated) styles (in serial or all-MPI parallelization with 1
task/core). This is because many of the USER-OMP styles contain
similar optimizations to those used in the OPT package, as described
above.
</P>
<P>With multiple threads/task, the optimal choice of MPI tasks/node and
OpenMP threads/task can vary a lot and should always be tested via
benchmark runs for a specific simulation running on a specific
machine, paying attention to guidelines discussed in the next
sub-section.
</P>
<P>A description of the multi-threading strategy used in the USER-OMP
package and some performance examples are <AHREF ="http://sites.google.com/site/akohlmey/software/lammps-icms/lammps-icms-tms2011-talk.pdf?attredirects=0&d=1">presented
here</A>
</P>
<P><B>Guidelines for best performance:</B>
</P>
<P>For many problems on current generation CPUs, running the USER-OMP
package with a single thread/task is faster than running with multiple
threads/task. This is because the MPI parallelization in LAMMPS is
often more efficient than multi-threading as implemented in the
USER-OMP package. The parallel efficiency (in a threaded sense) also
varies for different USER-OMP styles.
</P>
<P>Using multiple threads/task can be more effective under the following
circumstances:
</P>
<UL><LI>Individual compute nodes have a significant number of CPU cores but
the CPU itself has limited memory bandwidth, e.g. for Intel Xeon 53xx
(Clovertown) and 54xx (Harpertown) quad core processors. Running one
MPI task per CPU core will result in significant performance
degradation, so that running with 4 or even only 2 MPI tasks per node
is faster. Running in hybrid MPI+OpenMP mode will reduce the
inter-node communication bandwidth contention in the same way, but
offers an additional speedup by utilizing the otherwise idle CPU
cores.
<LI>The interconnect used for MPI communication does not provide
sufficient bandwidth for a large number of MPI tasks per node. For
example, this applies to running over gigabit ethernet or on Cray XT4
or XT5 series supercomputers. As in the aforementioned case, this
effect worsens when using an increasing number of nodes.
<LI>The system has a spatially inhomogeneous particle density which does
not map well to the <AHREF ="processors.html">domain decomposition scheme</A> or
<AHREF ="balance.html">load-balancing</A> options that LAMMPS provides. This is
because multi-threading achives parallelism over the number of
particles, not via their distribution in space.
<LI>A machine is being used in "capability mode", i.e. near the point
where MPI parallelism is maxed out. For example, this can happen when
using the <AHREF ="kspace_style.html">PPPM solver</A> for long-range
electrostatics on large numbers of nodes. The scaling of the KSpace
calculation (see the <AHREF ="kspace_style.html">kspace_style</A> command) becomes
the performance-limiting factor. Using multi-threading allows less
MPI tasks to be invoked and can speed-up the long-range solver, while
increasing overall performance by parallelizing the pairwise and
bonded calculations via OpenMP. Likewise additional speedup can be
sometimes be achived by increasing the length of the Coulombic cutoff
and thus reducing the work done by the long-range solver. Using the
<AHREF ="run_style.html">run_style verlet/split</A> command, which is compatible
with the USER-OMP package, is an alternative way to reduce the number
of MPI tasks assigned to the KSpace calculation.
</UL>
<P>Additional performance tips are as follows:
</P>
<UL><LI>The best parallel efficiency from <I>omp</I> styles is typically achieved
when there is at least one MPI task per physical processor,
i.e. socket or die.
<LI>It is usually most efficient to restrict threading to a single
socket, i.e. use one or more MPI task per socket.
<LI>Several current MPI implementation by default use a processor affinity
setting that restricts each MPI task to a single CPU core. Using
multi-threading in this mode will force the threads to share that core
and thus is likely to be counterproductive. Instead, binding MPI
tasks to a (multi-core) socket, should solve this issue.
