Various updates.
This commit was SVN r10404.
Этот коммит содержится в:
Jeff Squyres
родитель
01913be4e6
Коммит
beceebdecd
78
README
78
README
@ -40,7 +40,7 @@ Thanks for your time.
|
||||
===========================================================================
|
||||
|
||||
The following abbreviated list of release notes applies to this code
|
||||
base as of this writing (27 Feb 2006):
|
||||
base as of this writing (17 Jun 2006):
|
||||
|
||||
- Open MPI includes support for a wide variety of supplemental
|
||||
hardware and software package. When configuring Open MPI, you may
|
||||
@ -93,11 +93,14 @@ base as of this writing (27 Feb 2006):
|
||||
inoperable (see notes about MCA parameters later in this file). In
|
||||
particular, some parameters have required options that must be
|
||||
included.
|
||||
- The "btl" parameter must include the "self" component, or Open MPI
|
||||
will not be able to deliver messages to the same rank as the
|
||||
sender.
|
||||
- The "btl_tcp_if_exclude" paramater must include "lo", or Open MPI
|
||||
will not be able to route MPI messages using the TCP BTL.
|
||||
- If specified, the "btl" parameter must include the "self"
|
||||
component, or Open MPI will not be able to deliver messages to the
|
||||
same rank as the sender. For example: "mpirun --mca btl tcp,self
|
||||
..."
|
||||
- If specified, the "btl_tcp_if_exclude" paramater must include the
|
||||
loopback device ("lo" on many Linux platforms), or Open MPI will
|
||||
not be able to route MPI messages using the TCP BTL. For example:
|
||||
"mpirun --mca btl_tcp_if_exclude lo,eth1 ..."
|
||||
|
||||
- Building shared libraries on AIX with the xlc compilers is only
|
||||
supported if you supply the following command line option to
|
||||
@ -165,10 +168,10 @@ base as of this writing (27 Feb 2006):
|
||||
during MPI_FINALIZE.
|
||||
|
||||
- Running on nodes with different endian and/or different datatype
|
||||
sizes within a single parallel job is supported as of 1.1. However,
|
||||
we do not properly resize data when datatypes differ in size (for
|
||||
example, sending a 4 byte MPI_LONG and receiving an 8 byte MPI_LONG
|
||||
will fail).
|
||||
sizes within a single parallel job is supported starting with Open
|
||||
MPI v1.1. However, Open MPI does not resize data when datatypes
|
||||
differ in size (for example, sending a 4 byte MPI_LONG and receiving
|
||||
an 8 byte MPI_LONG will fail).
|
||||
|
||||
- MPI_THREAD_MULTIPLE support is included, but is only lightly tested.
|
||||
|
||||
@ -213,7 +216,7 @@ base as of this writing (27 Feb 2006):
|
||||
eliminate the extra overhead of software MPI message matching where
|
||||
possible.
|
||||
|
||||
- The Fortran 90 MPI bindings can now be built in one of four sizes
|
||||
- The Fortran 90 MPI bindings can now be built in one of three sizes
|
||||
using --with-mpi-f90-size=SIZE (see description below). These sizes
|
||||
reflect the number of MPI functions included in the "mpi" Fortran 90
|
||||
module and therefore which functions will be subject to strict type
|
||||
@ -239,22 +242,20 @@ base as of this writing (27 Feb 2006):
|
||||
dimensions specified by --with-f90-max-array-dim (default value is
|
||||
4).
|
||||
|
||||
- large: All MPI functions (i.e., all the functions in "medium" plus
|
||||
all MPI functions that take two choice buffers, such as
|
||||
MPI_SCATTER, MPI_GATHER, etc.). All the two-choice-buffer
|
||||
functions will have variants for each of the MPI-supported Fortran
|
||||
intrinsic types up to the number of dimensions specified by
|
||||
--with-f90-max-array-dim, but both buffers will be of the same
|
||||
type.
|
||||
|
||||
Increasing the size of the F90 module (in order from trivial, small,
|
||||
medium, and large) will generally increase the length of time
|
||||
required to compile user MPI applications. Specifically, "trivial"-
|
||||
and "small"-sized F90 modules generally allow user MPI applications
|
||||
to be compiled fairly quickly but lose type safety for all MPI
|
||||
functions with choice buffers. "medium"- and "large"-sized F90
|
||||
modules generally take longer to compile user applications but
|
||||
provide greater type safety for MPI functions.
|
||||
and medium) will generally increase the length of time required to
|
||||
compile user MPI applications. Specifically, "trivial"- and
|
||||
"small"-sized F90 modules generally allow user MPI applications to
|
||||
be compiled fairly quickly but lose type safety for all MPI
|
||||
functions with choice buffers. "medium"-sized F90 modules generally
|
||||
take longer to compile user applications but provide greater type
|
||||
safety for MPI functions.
|
||||
|
||||
Note that MPI functions with two choice buffers (e.g., MPI_GATHER)
|
||||
are not currently included in Open MPI's F90 interface. Calls to
|
||||
these functions will automatically fall through to Open MPI's F77
|
||||
interface. A "large" size that includes the two choice buffer MPI
|
||||
functions is possible in future versions of Open MPI.
|
||||
|
||||
===========================================================================
|
||||
|
||||
@ -290,8 +291,11 @@ for a full list); a summary of the more commonly used ones follows:
|
||||
located. This enables mVAPI support in Open MPI.
|
||||
|
||||
--with-openib=<directory>
|
||||
Specify the directory where the Open IB libraries and header files are
|
||||
located. This enables mVAPI support in Open MPI.
|
||||
Specify the directory where the Open Fabrics (previously known as
|
||||
OpenIB) libraries and header files are located. This enables Open
|
||||
Fabrics support in Open MPI. This option will likely be be
|
||||
deprecated in favor of "--with-openfrabrics" in a future version of
|
||||
Open MPI.
|
||||
|
||||
--with-tm=<directory>
|
||||
Specify the directory where the TM libraries and header files are
|
||||
@ -339,13 +343,11 @@ for a full list); a summary of the more commonly used ones follows:
|
||||
--with-f90-max-array-dim and --with-mpi-f90-size options.
|
||||
|
||||
--with-mpi-f90-size=<SIZE>
|
||||
Four sizes of the MPI F90 module can be built: trivial (only a
|
||||
handful of MPI-2 F90-specific functions are included in the F90
|
||||
Three sizes of the MPI F90 module can be built: trivial (only a
|
||||
handful of MPI-2 F90-specific functions are included in the F90
|
||||
module), small (trivial + all MPI functions that take no choice
|
||||
buffers), medium (small + all MPI functions that take 1 choice
|
||||
buffer), and large (medium + all MPI functions that take 2 choice
|
||||
buffers, but only where the types of both choice buffers are the
|
||||
same). This parameter is only used if the F90 bindings are
|
||||
buffers), and medium (small + all MPI functions that take 1 choice
|
||||
buffer). This parameter is only used if the F90 bindings are
|
||||
enabled.
|
||||
|
||||
--with-f90-max-array-dim=<DIM>
|
||||
@ -358,11 +360,15 @@ for a full list); a summary of the more commonly used ones follows:
|
||||
By default, libmpi is built as a shared library, and all components
|
||||
are built as dynamic shared objects (DSOs). This switch disables
|
||||
this default; it is really only useful when used with
|
||||
--enable-static.
|
||||
--enable-static. Specifically, this option does *not* imply
|
||||
--disable-shared; enabling static libraries and disabling shared
|
||||
libraries are two independent options.
|
||||
|
||||
--enable-static
|
||||
Build libmpi as a static library, and statically link in all
|
||||
components.
|
||||
components. Note that this option does *not* imply
|
||||
--disable-shared; enabling static libraries and disabling shared
|
||||
libraries are two independent options.
|
||||
|
||||
There are several other options available -- see "./configure --help".
|
||||
|
||||
|
||||
Загрузка…
Ссылка в новой задаче
Block a user