1
1
openmpi/opal/mca/btl/usnic/configure.m4

146 строки
5.9 KiB
Plaintext
Исходник Обычный вид История

# -*- shell-script -*-
#
# Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
# University Research and Technology
# Corporation. All rights reserved.
# Copyright (c) 2004-2005 The University of Tennessee and The University
# of Tennessee Research Foundation. All rights
# reserved.
# Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
# University of Stuttgart. All rights reserved.
# Copyright (c) 2004-2005 The Regents of the University of California.
# All rights reserved.
# Copyright (c) 2006 Sandia National Laboratories. All rights
# reserved.
# Copyright (c) 2010-2015 Cisco Systems, Inc. All rights reserved.
# $COPYRIGHT$
#
# 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 00:47:28 +00:00
# MCA_opal_btl_usnic_POST_CONFIG([should_build])
# ------------------------------------------
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 00:47:28 +00:00
AC_DEFUN([MCA_opal_btl_usnic_POST_CONFIG], [
AM_CONDITIONAL([OPAL_BTL_USNIC_BUILD_UNIT_TESTS],
[test "$1" -eq 1 && test "X$enable_opal_btl_usnic_unit_tests" = "Xyes"])
])
# MCA_btl_usnic_CONFIG([action-if-can-copalle],
# [action-if-cant-copalle])
# ------------------------------------------------
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 00:47:28 +00:00
AC_DEFUN([MCA_opal_btl_usnic_CONFIG],[
AC_CONFIG_FILES([opal/mca/btl/usnic/Makefile])
AC_ARG_WITH([usnic],
[AS_HELP_STRING([--with-usnic],
[If specified, cause an error if usNIC
support cannot be built])])
# If --without-usnic was specified, then gracefully exit.
# Otherwise, do the rest of the config.
AS_IF([test "x$with_usnic" = "xno"],
[AC_MSG_WARN([--without-usnic specified; skipping usnic BTL])
$2],
[_OPAL_BTL_USNIC_DO_CONFIG($1, $2)])
])
AC_DEFUN([_OPAL_BTL_USNIC_DO_CONFIG],[
OPAL_VAR_SCOPE_PUSH([unit_tests])
# see README.test for information about this scheme
AC_ARG_ENABLE([opal-btl-usnic-unit-tests],
[AS_HELP_STRING([--enable-opal-btl-usnic-unit-tests],
[build unit tests for the usnic BTL,
including the test runner program,
George did the work and deserves all the credit for it. Ralph did the merge, and deserves whatever blame results from errors in it :-) WHAT: Open our low-level communication infrastructure by moving all necessary components (btl/rcache/allocator/mpool) down in OPAL All the components required for inter-process communications are currently deeply integrated in the OMPI layer. Several groups/institutions have express interest in having a more generic communication infrastructure, without all the OMPI layer dependencies. This communication layer should be made available at a different software level, available to all layers in the Open MPI software stack. As an example, our ORTE layer could replace the current OOB and instead use the BTL directly, gaining access to more reactive network interfaces than TCP. Similarly, external software libraries could take advantage of our highly optimized AM (active message) communication layer for their own purpose. UTK with support from Sandia, developped a version of Open MPI where the entire communication infrastucture has been moved down to OPAL (btl/rcache/allocator/mpool). Most of the moved components have been updated to match the new schema, with few exceptions (mainly BTLs where I have no way of compiling/testing them). Thus, the completion of this RFC is tied to being able to completing this move for all BTLs. For this we need help from the rest of the Open MPI community, especially those supporting some of the BTLs. A non-exhaustive list of BTLs that qualify here is: mx, portals4, scif, udapl, ugni, usnic. This commit was SVN r32317.
2014-07-26 00:47:28 +00:00
opal_btl_usnic_run_tests])])
AS_IF([test "X$enable_opal_btl_usnic_unit_tests" = "Xyes"],
[unit_tests=1
AC_MSG_NOTICE([enabling usnic BTL unit tests])],
[unit_tests=0])
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 00:47:28 +00:00
AC_DEFINE_UNQUOTED([OPAL_BTL_USNIC_UNIT_TESTS], [$unit_tests],
[define to 1 if usnic BTL unit tests are enabled, 0 otherwise])
unset unit_tests
# The current logic in btl_usnic_compat.h checks the OPAL version as a
# proxy for the top-level OPAL version. Unfortunately this does the wrong
# thing for other top-level projects that might use the usnic BTL, such as
# ORCM. ORCM's versioning is totally unrelated to OPAL's. As a short term
# workaround, just disqualify ourselves if the OPAL version seems too old.
# In the longer term we should be doing something else, like versioning
# OPAL and OPAL separately.
opal_btl_usnic_happy=yes
AS_IF([test "$OPAL_MAJOR_VERSION" -eq "1" && \
test "$OPAL_MINOR_VERSION" -lt "7"],
[AC_MSG_NOTICE([OPAL version appears to be too old, disabling the usnic BTL])
opal_btl_usnic_happy=no])
# We only want to build on 64 bit Linux.
AS_IF([test "$opal_btl_usnic_happy" = "yes"],
[AC_CHECK_SIZEOF([void *])
AC_MSG_CHECKING([for 64 bit Linux])
case $host_os in
*linux*)
AS_IF([test $ac_cv_sizeof_void_p -eq 8],
[],
[opal_btl_usnic_happy=no])
;;
*)
opal_btl_usnic_happy=no
;;
esac
AC_MSG_RESULT([$opal_btl_usnic_happy])
])
# The usnic BTL requires libfabric support.
AS_IF([test "$opal_btl_usnic_happy" = "yes"],
[AC_MSG_CHECKING([whether libfabric support is available])
AS_IF([test "$opal_common_libfabric_happy" = "yes"],
[opal_btl_usnic_happy=yes],
[opal_btl_usnic_happy=no])
AC_MSG_RESULT([$opal_btl_usnic_happy])
])
# The usnic BTL requires at least libfabric v1.1 (there was a
# critical bug in libfabric v1.0).
AS_IF([test "$opal_btl_usnic_happy" = "yes"],
[AC_MSG_CHECKING([whether libfabric is >= v1.1])
opal_btl_usnic_CPPFLAGS_save=$CPPFLAGS
CPPFLAGS="$opal_common_libfabric_CPPFLAGS $CPPFLAGS"
AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <rdma/fabric.h>]],
[[
#if !defined(FI_MAJOR_VERSION)
#error your version of libfabric is too old
#elif FI_VERSION(FI_MAJOR_VERSION, FI_MINOR_VERSION) < FI_VERSION(1, 1)
#error your version of libfabric is too old
#endif
]])],
[opal_btl_usnic_happy=yes],
[opal_btl_usnic_happy=no])
AC_MSG_RESULT([$opal_btl_usnic_happy])
CPPFLAGS=$opal_btl_usnic_CPPFLAGS_save
])
# Make sure we can find the libfabric usnic extensions header
AS_IF([test "$opal_btl_usnic_happy" = "yes" ],
[opal_btl_usnic_CPPFLAGS_save=$CPPFLAGS
CPPFLAGS="$opal_common_libfabric_CPPFLAGS $CPPFLAGS"
AC_CHECK_HEADER([rdma/fi_ext_usnic.h],
[],
[opal_btl_usnic_happy=no])
CPPFLAGS=$opal_btl_usnic_CPPFLAGS_save
])
# All done
AS_IF([test "$opal_btl_usnic_happy" = "yes"],
[$1],
[AS_IF([test "$with_usnic" = "yes"],
[AC_MSG_WARN([--with-usnic was specified, but Cisco usNIC support cannot be built])
AC_MSG_ERROR([Cannot continue])],
[$2])
])
OPAL_VAR_SCOPE_POP
])dnl