jianmu
/
lammps
forked from lijiext/lammps
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

git-svn-id: svn://svn.icms.temple.edu/lammps-ro/trunk@7348 f3b2605a-c512-4ea7-a41b-209d697bcdaa

This commit is contained in:
sjplimp 2011-12-13 18:30:06 +00:00
parent 2f7c6cfb0d
commit eeb5ad77e6
2 changed files with 50 additions and 42 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

46
doc/processors.html
View File

@ -28,8 +28,8 @@
Nc = number of cores per node Nc = number of cores per node
Cx,Cy,Cz = # of cores in each dimension of 3d sub-grid assigned to each node Cx,Cy,Cz = # of cores in each dimension of 3d sub-grid assigned to each node
numa params = none numa params = none
custom params = inname custom params = infile
inname = file containing grid layout infile = file containing grid layout
<I>map</I> arg = <I>cart</I> or <I>cart/reorder</I> or <I>xyz</I> or <I>xzy</I> or <I>yxz</I> or <I>yzx</I> or <I>zxy</I> or <I>zyx</I> <I>map</I> arg = <I>cart</I> or <I>cart/reorder</I> or <I>xyz</I> or <I>xzy</I> or <I>yxz</I> or <I>yzx</I> or <I>zxy</I> or <I>zyx</I>
cart = use MPI_Cart() methods to map processors to 3d grid with reorder = 0 cart = use MPI_Cart() methods to map processors to 3d grid with reorder = 0
cart/reorder = use MPI_Cart() methods to map processors to 3d grid with reorder = 1 cart/reorder = use MPI_Cart() methods to map processors to 3d grid with reorder = 1
@ -40,8 +40,8 @@
Precv = partition # (1 to Np) which will recv the processor layout Precv = partition # (1 to Np) which will recv the processor layout
cstyle = <I>multiple</I> cstyle = <I>multiple</I>
<I>multiple</I> = Psend grid will be multiple of Precv grid in each dimension <I>multiple</I> = Psend grid will be multiple of Precv grid in each dimension
<I>file</I> arg = outname <I>file</I> arg = outfile
outname = name of file to write 3d grid of processors to outfile = name of file to write 3d grid of processors to
</PRE> </PRE>
</UL> </UL>
@ -112,12 +112,12 @@ Px,Py,Pz settings, and which minimizes the surface-to-volume ratio of
each processor's sub-domain, as described above. The mapping of each processor's sub-domain, as described above. The mapping of
processors to the grid is determined by the <I>map</I> keyword setting. processors to the grid is determined by the <I>map</I> keyword setting.
</P> </P>
<P>The <I>twolevel</I> style can be used on machines with multi-core nodes <P>The <I>twolevel</I> style can be used on machines with multicore nodes to
to minimize off-node communication. It insures that contiguous minimize off-node communication. It insures that contiguous
sub-sections of the 3d grid are assigned to all the cores of a node. sub-sections of the 3d grid are assigned to all the cores of a node.
For example if <I>Nc</I> is 4, then 2x2x1 or 2x1x2 or 1x2x2 sub-sections For example if <I>Nc</I> is 4, then 2x2x1 or 2x1x2 or 1x2x2 sub-sections of
of the 3d grid will correspond to the cores of each node. This the 3d grid will correspond to the cores of each node. This affects
affects both the factorization and mapping steps. both the factorization and mapping steps.
</P> </P>
<P>The <I>Cx</I>, <I>Cy</I>, <I>Cz</I> settings are similar to the <I>Px</I>, <I>Py</I>, <I>Pz</I> <P>The <I>Cx</I>, <I>Cy</I>, <I>Cz</I> settings are similar to the <I>Px</I>, <I>Py</I>, <I>Pz</I>
settings, only their product should equal <I>Nc</I>. Any of the 3 settings, only their product should equal <I>Nc</I>. Any of the 3
@ -156,7 +156,7 @@ any of the Px or Py of Pz values is greater than 1.
the MPI ranks of processors LAMMPS is running on are ordered by core the MPI ranks of processors LAMMPS is running on are ordered by core
and then by node. See the same note for the <I>twolevel</I> keyword. and then by node. See the same note for the <I>twolevel</I> keyword.
</P> </P>
<P>The <I>custom</I> style uses the file <I>inname</I> to define both the 3d <P>The <I>custom</I> style uses the file <I>infile</I> to define both the 3d
factorization and the mapping of processors to the grid. factorization and the mapping of processors to the grid.
</P> </P>
<P>The file should have the following format. Any number of initial <P>The file should have the following format. Any number of initial
@ -268,11 +268,11 @@ setup phase if this error has been made.
</P> </P>
<HR> <HR>
<P>The <I>out</I> keyword writes the mapping of the factorization of P <P>The <I>file</I> keyword writes the mapping of the factorization of P
processors and their mapping to the 3d grid to the specified file processors and their mapping to the 3d grid to the specified file
<I>fname</I>. This is useful to check that you assigned physical <I>outfile</I>. This is useful to check that you assigned physical
processors in the manner you desired, which can be tricky to figure processors in the manner you desired, which can be tricky to figure
out, especially when running on multiple partitions or on a multicore out, especially when running on multiple partitions or on, a multicore
machine or when the processor ranks were reordered by use of the machine or when the processor ranks were reordered by use of the
<A HREF = "Section_start.html#start_6">-reorder command-line switch</A> or due to <A HREF = "Section_start.html#start_6">-reorder command-line switch</A> or due to
use of MPI-specific launch options such as a config file. use of MPI-specific launch options such as a config file.
@ -282,10 +282,9 @@ to a different file, e.g. using a <A HREF = "variable.html">world-style variable
for the filename. The file has a self-explanatory header, followed by for the filename. The file has a self-explanatory header, followed by
one-line per processor in this format: one-line per processor in this format:
</P> </P>
<P>I J K: world-ID universe-ID original-ID: name <P>world-ID universe-ID original-ID: I J K: name
</P> </P>
<P>I,J,K are the indices of the processor in the 3d logical grid. The <P>The IDs are the processor's rank in this simulation (the world), the
IDs are the processor's rank in this simulation (the world), the
universe (of multiple simulations), and the original MPI communicator universe (of multiple simulations), and the original MPI communicator
used to instantiate LAMMPS, respectively. The world and universe IDs used to instantiate LAMMPS, respectively. The world and universe IDs
will only be different if you are running on more than one partition; will only be different if you are running on more than one partition;
@ -293,11 +292,16 @@ see the <A HREF = "Section_start.html#start_6">-partition command-line switch</A
The universe and original IDs will only be different if you used the The universe and original IDs will only be different if you used the
<A HREF = "Section_start.html#start_6">-reorder command-line switch</A> to reorder <A HREF = "Section_start.html#start_6">-reorder command-line switch</A> to reorder
the processors differently than their rank in the original the processors differently than their rank in the original
communicator LAMMPS was instantiated with. The <I>name</I> is what is communicator LAMMPS was instantiated with.
returned by a call to MPI_Get_processor_name() and should represent an </P>
identifier relevant to the physical processors in your machine. Note <P>I,J,K are the indices of the processor in the 3d logical grid, each
that depending on the MPI implementation, multiple cores can have the from 1 to Nd, where Nd is the number of processors in that dimension
same <I>name</I>. of the grid.
</P>
<P>The <I>name</I> is what is returned by a call to MPI_Get_processor_name()
and should represent an identifier relevant to the physical processors
in your machine. Note that depending on the MPI implementation,
multiple cores can have the same <I>name</I>.
</P> </P>
<HR> <HR>

46
doc/processors.txt
View File

@ -22,8 +22,8 @@ keyword = {grid} or {map} or {part} or {file} :l
Nc = number of cores per node Nc = number of cores per node
Cx,Cy,Cz = # of cores in each dimension of 3d sub-grid assigned to each node Cx,Cy,Cz = # of cores in each dimension of 3d sub-grid assigned to each node
numa params = none numa params = none
custom params = inname custom params = infile
inname = file containing grid layout infile = file containing grid layout
{map} arg = {cart} or {cart/reorder} or {xyz} or {xzy} or {yxz} or {yzx} or {zxy} or {zyx} {map} arg = {cart} or {cart/reorder} or {xyz} or {xzy} or {yxz} or {yzx} or {zxy} or {zyx}
cart = use MPI_Cart() methods to map processors to 3d grid with reorder = 0 cart = use MPI_Cart() methods to map processors to 3d grid with reorder = 0
cart/reorder = use MPI_Cart() methods to map processors to 3d grid with reorder = 1 cart/reorder = use MPI_Cart() methods to map processors to 3d grid with reorder = 1
@ -34,8 +34,8 @@ keyword = {grid} or {map} or {part} or {file} :l
Precv = partition # (1 to Np) which will recv the processor layout Precv = partition # (1 to Np) which will recv the processor layout
cstyle = {multiple} cstyle = {multiple}
{multiple} = Psend grid will be multiple of Precv grid in each dimension {multiple} = Psend grid will be multiple of Precv grid in each dimension
{file} arg = outname {file} arg = outfile
outname = name of file to write 3d grid of processors to :pre outfile = name of file to write 3d grid of processors to :pre
:ule :ule
[Examples:] [Examples:]
@ -105,12 +105,12 @@ Px,Py,Pz settings, and which minimizes the surface-to-volume ratio of
each processor's sub-domain, as described above. The mapping of each processor's sub-domain, as described above. The mapping of
processors to the grid is determined by the {map} keyword setting. processors to the grid is determined by the {map} keyword setting.
The {twolevel} style can be used on machines with multi-core nodes The {twolevel} style can be used on machines with multicore nodes to
to minimize off-node communication. It insures that contiguous minimize off-node communication. It insures that contiguous
sub-sections of the 3d grid are assigned to all the cores of a node. sub-sections of the 3d grid are assigned to all the cores of a node.
For example if {Nc} is 4, then 2x2x1 or 2x1x2 or 1x2x2 sub-sections For example if {Nc} is 4, then 2x2x1 or 2x1x2 or 1x2x2 sub-sections of
of the 3d grid will correspond to the cores of each node. This the 3d grid will correspond to the cores of each node. This affects
affects both the factorization and mapping steps. both the factorization and mapping steps.
The {Cx}, {Cy}, {Cz} settings are similar to the {Px}, {Py}, {Pz} The {Cx}, {Cy}, {Cz} settings are similar to the {Px}, {Py}, {Pz}
settings, only their product should equal {Nc}. Any of the 3 settings, only their product should equal {Nc}. Any of the 3
@ -149,7 +149,7 @@ IMPORTANT NOTE: For the {numa} style to work correctly, it assumes
the MPI ranks of processors LAMMPS is running on are ordered by core the MPI ranks of processors LAMMPS is running on are ordered by core
and then by node. See the same note for the {twolevel} keyword. and then by node. See the same note for the {twolevel} keyword.
The {custom} style uses the file {inname} to define both the 3d The {custom} style uses the file {infile} to define both the 3d
factorization and the mapping of processors to the grid. factorization and the mapping of processors to the grid.
The file should have the following format. Any number of initial The file should have the following format. Any number of initial
@ -261,11 +261,11 @@ setup phase if this error has been made.
:line :line
The {out} keyword writes the mapping of the factorization of P The {file} keyword writes the mapping of the factorization of P
processors and their mapping to the 3d grid to the specified file processors and their mapping to the 3d grid to the specified file
{fname}. This is useful to check that you assigned physical {outfile}. This is useful to check that you assigned physical
processors in the manner you desired, which can be tricky to figure processors in the manner you desired, which can be tricky to figure
out, especially when running on multiple partitions or on a multicore out, especially when running on multiple partitions or on, a multicore
machine or when the processor ranks were reordered by use of the machine or when the processor ranks were reordered by use of the
"-reorder command-line switch"_Section_start.html#start_6 or due to "-reorder command-line switch"_Section_start.html#start_6 or due to
use of MPI-specific launch options such as a config file. use of MPI-specific launch options such as a config file.
@ -275,10 +275,9 @@ to a different file, e.g. using a "world-style variable"_variable.html
for the filename. The file has a self-explanatory header, followed by for the filename. The file has a self-explanatory header, followed by
one-line per processor in this format: one-line per processor in this format:
I J K: world-ID universe-ID original-ID: name world-ID universe-ID original-ID: I J K: name
I,J,K are the indices of the processor in the 3d logical grid. The The IDs are the processor's rank in this simulation (the world), the
IDs are the processor's rank in this simulation (the world), the
universe (of multiple simulations), and the original MPI communicator universe (of multiple simulations), and the original MPI communicator
used to instantiate LAMMPS, respectively. The world and universe IDs used to instantiate LAMMPS, respectively. The world and universe IDs
will only be different if you are running on more than one partition; will only be different if you are running on more than one partition;
@ -286,11 +285,16 @@ see the "-partition command-line switch"_Section_start.html#start_6.
The universe and original IDs will only be different if you used the The universe and original IDs will only be different if you used the
"-reorder command-line switch"_Section_start.html#start_6 to reorder "-reorder command-line switch"_Section_start.html#start_6 to reorder
the processors differently than their rank in the original the processors differently than their rank in the original
communicator LAMMPS was instantiated with. The {name} is what is communicator LAMMPS was instantiated with.
returned by a call to MPI_Get_processor_name() and should represent an
identifier relevant to the physical processors in your machine. Note I,J,K are the indices of the processor in the 3d logical grid, each
that depending on the MPI implementation, multiple cores can have the from 1 to Nd, where Nd is the number of processors in that dimension
same {name}. of the grid.
The {name} is what is returned by a call to MPI_Get_processor_name()
and should represent an identifier relevant to the physical processors
in your machine. Note that depending on the MPI implementation,
multiple cores can have the same {name}.
:line :line