2011-10-21 01:39:44 +04:00
|
|
|
# -*- shell-script -*-
|
|
|
|
|
#
|
|
|
|
|
# Copyright (c) 2009 The University of Tennessee and The University
|
|
|
|
|
# of Tennessee Research Foundation. All rights
|
|
|
|
|
# reserved.
|
|
|
|
|
# Copyright (c) 2009-2010 Cisco Systems, Inc. All rights reserved.
|
2014-01-29 22:35:47 +04:00
|
|
|
# Copyright (c) 2011-2014 Los Alamos National Security, LLC. All rights
|
|
|
|
|
# reserved.
|
2011-10-21 01:39:44 +04:00
|
|
|
# $COPYRIGHT$
|
|
|
|
|
#
|
|
|
|
|
# Additional copyrights may follow
|
|
|
|
|
#
|
|
|
|
|
# $HEADER$
|
|
|
|
|
#
|
|
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
# OPAL_CHECK_XPMEM(prefix, [action-if-found], [action-if-not-found])
|
2011-10-21 01:39:44 +04:00
|
|
|
# --------------------------------------------------------
|
|
|
|
|
# check if XPMEM support can be found. sets prefix_{CPPFLAGS,
|
|
|
|
|
# LDFLAGS, LIBS} as needed and runs action-if-found if there is
|
|
|
|
|
# support, otherwise executes action-if-not-found
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
AC_DEFUN([OPAL_CHECK_XPMEM], [
|
|
|
|
|
OPAL_VAR_SCOPE_PUSH([opal_check_xpmem_happy])
|
2011-10-21 01:39:44 +04:00
|
|
|
AC_ARG_WITH([xpmem],
|
|
|
|
|
[AC_HELP_STRING([--with-xpmem(=DIR)],
|
2011-10-21 19:45:05 +04:00
|
|
|
[Build with XPMEM kernel module support, searching for headers in DIR])])
|
2014-05-01 19:38:07 +04:00
|
|
|
OPAL_CHECK_WITHDIR([xpmem], [$with_xpmem], [include/xpmem.h include/sn/xpmem.h])
|
2011-10-21 01:39:44 +04:00
|
|
|
|
|
|
|
|
AC_ARG_WITH([xpmem-libdir],
|
|
|
|
|
[AC_HELP_STRING([--with-xpmem-libdir=DIR],
|
|
|
|
|
[Search for XPMEM library in DIR])])
|
2014-05-01 19:38:07 +04:00
|
|
|
OPAL_CHECK_WITHDIR([xpmem-libdir], [$with_xpmem_libdir], [libxpmem.*])
|
2011-10-21 01:39:44 +04:00
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
opal_check_xpmem_happy="no"
|
2014-01-29 22:35:47 +04:00
|
|
|
|
|
|
|
|
if test ! "$with_xpmem" = "no" ; then
|
|
|
|
|
if test ! -z "$with_xpmem" -a "$with_xpmem" != "yes" ; then
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
opal_check_xpmem_dir="$with_xpmem"
|
2014-01-29 22:35:47 +04:00
|
|
|
fi
|
|
|
|
|
|
|
|
|
|
if test ! -z "$with_xpmem_libdir" -a "$with_xpmem_libdir" != "yes" ; then
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
opal_check_xpmem_libdir="$with_xpmem_libdir"
|
2014-01-29 22:35:47 +04:00
|
|
|
fi
|
|
|
|
|
|
2014-05-01 18:24:56 +04:00
|
|
|
OPAL_CHECK_PACKAGE([$1],[xpmem.h sn/xpmem.h],[xpmem],[xpmem_make],[],
|
2014-07-29 22:59:59 +04:00
|
|
|
[$opal_check_xpmem_dir],[$opal_check_xpmem_libdir], [opal_check_xpmem_happy="yes"], [])
|
2014-01-29 22:35:47 +04:00
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
if test "$opal_check_xpmem_happy" = "no" -a -n "$with_xpmem" -a "$with_xpmem" != "yes" ; then
|
2014-01-29 22:35:47 +04:00
|
|
|
AC_MSG_ERROR([XPMEM support requested but not found. Aborting])
|
|
|
|
|
fi
|
|
|
|
|
fi
|
|
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
AS_IF([test "$opal_check_xpmem_happy" = "yes"], [$2], [$3])
|
2014-01-29 22:35:47 +04:00
|
|
|
|
2014-01-06 23:51:44 +04:00
|
|
|
OPAL_VAR_SCOPE_POP
|
2011-10-21 01:39:44 +04:00
|
|
|
])dnl
|
|
|
|
|
|
|
|
|
|
# MCA_btl_sm_CONFIG([action-if-can-compile],
|
|
|
|
|
# [action-if-cant-compile])
|
|
|
|
|
# ------------------------------------------------
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
AC_DEFUN([MCA_opal_btl_vader_CONFIG],[
|
|
|
|
|
AC_CONFIG_FILES([opal/mca/btl/vader/Makefile])
|
2011-10-21 01:39:44 +04:00
|
|
|
|
2014-01-06 23:51:44 +04:00
|
|
|
OPAL_VAR_SCOPE_PUSH([btl_vader_xpmem_happy btl_vader_cma_happy])
|
|
|
|
|
|
|
|
|
|
btl_vader_cma_happy=0
|
|
|
|
|
btl_vader_xpmem_happy=0
|
|
|
|
|
|
|
|
|
|
# default to using XPMEM if it is available
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
OPAL_CHECK_XPMEM([btl_vader], [btl_vader_xpmem_happy=1], [])
|
2014-01-06 23:51:44 +04:00
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
AC_DEFINE_UNQUOTED([OPAL_BTL_VADER_HAVE_XPMEM], [$btl_vader_xpmem_happy],
|
2014-01-06 23:51:44 +04:00
|
|
|
[If XPMEM support can be enabled within vader])
|
2011-10-21 01:39:44 +04:00
|
|
|
|
2014-01-06 23:51:44 +04:00
|
|
|
if test $btl_vader_xpmem_happy = 0 ; then
|
|
|
|
|
# check for CMA if requested. it won't be used if xpmem was available
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
OPAL_CHECK_CMA([btl_vader], [btl_vader_cma_happy=1], [])
|
2014-01-06 23:51:44 +04:00
|
|
|
fi
|
2011-10-21 19:45:05 +04:00
|
|
|
|
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-)
WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL
All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic.
This commit was SVN r32317.
2014-07-26 04:47:28 +04:00
|
|
|
AC_DEFINE_UNQUOTED([OPAL_BTL_VADER_HAVE_CMA], [$btl_vader_cma_happy],
|
2014-01-06 23:51:44 +04:00
|
|
|
[If CMA support can be enabled within vader])
|
|
|
|
|
|
|
|
|
|
OPAL_VAR_SCOPE_POP
|
|
|
|
|
|
|
|
|
|
# always happy
|
|
|
|
|
[$1]
|
2011-10-21 01:39:44 +04:00
|
|
|
|
|
|
|
|
# substitute in the things needed to build with XPMEM support
|
|
|
|
|
AC_SUBST([btl_vader_CFLAGS])
|
|
|
|
|
AC_SUBST([btl_vader_CPPFLAGS])
|
|
|
|
|
AC_SUBST([btl_vader_LDFLAGS])
|
|
|
|
|
AC_SUBST([btl_vader_LIBS])
|
|
|
|
|
])dnl
|