ICODE-ADC/lammps
4
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update docs

Browse Source
This commit is contained in:
Stan Gerald Moore
2024-02-14 16:50:52 -07:00
parent 52aefdd117
commit 6ebc32fa29
1 changed files with 20 additions and 12 deletions
Download Patch File Download Diff File Expand all files Collapse all files

32
doc/src/package.rst
View File

@ -74,7 +74,7 @@ Syntax
*no_affinity* values = none
*kokkos* args = keyword value ...
zero or more keyword/value pairs may be appended
keywords = *neigh* or *neigh/qeq* or *neigh/thread* or *neigh/transpose* or *newton* or *binsize* or *comm* or *comm/exchange* or *comm/forward* or *comm/pair/forward* or *comm/fix/forward* or *comm/reverse* or *comm/pair/reverse* or *sort* or *gpu/aware* or *pair/only*
keywords = *neigh* or *neigh/qeq* or *neigh/thread* or *neigh/transpose* or *newton* or *binsize* or *comm* or *comm/exchange* or *comm/forward* or *comm/pair/forward* or *comm/fix/forward* or *comm/reverse* or *comm/pair/reverse* or *sort* or *atom/map* or *gpu/aware* or *pair/only*
*neigh* value = *full* or *half*
full = full neighbor list
half = half neighbor list built in thread-safe manner
@ -108,6 +108,9 @@ Syntax
*sort* value = *no* or *device*
*no* = perform atom sorting in non-KOKKOS mode
*device* = perform atom sorting on device (e.g. on GPU)
*atom/map* value = *no* or *device*
*no* = build atom map in non-KOKKOS mode
*device* = build atom map on device (e.g. on GPU)
*gpu/aware* = *off* or *on*
*off* = do not use GPU-aware MPI
*on* = use GPU-aware MPI (default)
@ -576,6 +579,11 @@ effect if the simulation is running on the CPU. Not all fix styles with
extra atom data support *device* mode and in that case a warning will be
given and atom sorting will run in *no* mode instead.
The *atom/map* keyword determines whether the host or device builds the
atom_map, see the :doc:`atom_modify map <atom_modify>` command. The
value options for the *atom/map* keyword are identical to the *sort*
keyword above.
The *gpu/aware* keyword chooses whether GPU-aware MPI will be used. When
this keyword is set to *on*, buffers in GPU memory are passed directly
through MPI send/receive calls. This reduces overhead of first copying
@ -724,17 +732,17 @@ made automatically if the "-sf intel" :doc:`command-line switch <Run_options>`
is used. If it is not used, you must invoke the package intel command in your
input script or via the "-pk intel" :doc:`command-line switch <Run_options>`.
For the KOKKOS package, the option defaults for GPUs are neigh = full,
neigh/qeq = full, newton = off, binsize for GPUs = 2x LAMMPS default value,
comm = device, sort = device, neigh/transpose = off, gpu/aware = on. When
LAMMPS can safely detect that GPU-aware MPI is not available, the default value
of gpu/aware becomes "off". For CPUs or Xeon Phis, the option defaults are
neigh = half, neigh/qeq = half, newton = on, binsize = 0.0, comm = no, and sort
= no. For GPUs, option neigh/thread = on when there are 16k atoms or less on an MPI
rank, otherwise it is "off". These settings are made automatically by the
required "-k on" :doc:`command-line switch <Run_options>`. You can change them
by using the package kokkos command in your input script or via the :doc:`-pk
kokkos command-line switch <Run_options>`.
For the KOKKOS package, the option defaults for GPUs are neigh = full, neigh/qeq
= full, newton = off, binsize for GPUs = 2x LAMMPS default value, comm = device,
sort = device, atom/map = device, neigh/transpose = off, and gpu/aware = on.
When LAMMPS can safely detect that GPU-aware MPI is not available, the default
value of gpu/aware becomes "off". For CPUs or Xeon Phis, the option defaults are
neigh = half, neigh/qeq = half, newton = on, binsize = 0.0, comm = no, sort =
no, and atom/map = no. For GPUs, option neigh/thread = on when there are 16k
atoms or less on an MPI rank, otherwise it is "off". These settings are made
automatically by the required "-k on" :doc:`command-line switch <Run_options>`.
You can change them by using the package kokkos command in your input script or
via the :doc:`-pk kokkos command-line switch <Run_options>`.
For the OMP package, the default is Nthreads = 0 and the option defaults are
neigh = yes. These settings are made automatically if the "-sf omp"