Intel MPI<\/p>\n
mpiicc thread_support.c -mt_mpi<\/strong><\/p>\n MVAPICH2<\/p>\n mpirun_rsh \u00c2\u00a0-np 2 \u00e2\u20ac\u201cenv MV2_USE_CUDA=1\u00c2\u00a0<\/strong>MV2_ENABLE_ Cary MPICH2<\/p>\n Fine-grained Multi-threading Support This feature requires CLE release 5.1 or higher, CrayPE version 2.2.0 At this time, this library supports only MPI point-to-point Support for larger subsets of MPI operations will be provided in To use the fine-grained threading feature, follow these steps.<\/p>\n Step 1. Compile a multi-threaded MPI application using the If the code uses static linking:<\/p>\n cc -craympich-mt -o mpi_mt_test.x mpi_mt_test.c<\/p>\n If the code uses dynamic linking:<\/p>\n cc -dynamic -craympich-mt -o mpi_mt_test.x mpi_mt_test.c<\/p>\n Note: The -craympich-mt and -default64 compiler command line Step 2. Set the environment variables.<\/p>\n export MPICH_MAX_THREAD_SAFETY=multiple<\/p>\n export MPICH_VERSION_DISPLAY=1<\/p>\n Step 3. Run the executable.<\/p>\n aprun arguments .\/executable_name<\/p>\n The MPICH_VERSION_DISPLAY environment variable is set in order to BUILD INFO : Built Date Time Year (svn rev rev) MT-G CPR-R<\/p>\n Where MT-G corresponds to the default Multi-threaded MPI library with The new fine-grained threading library displays this version string:<\/p>\n BUILD INFO : Built Date Time Year (svn rev rev) MT-O CPR-R<\/p>\n Where MT-O corresponds to the fine-grained Multi-threaded MPI library Intel MPI mpiicc thread_support.c -mt_mpi MVAPICH2 mpirun_rsh \u00c2\u00a0-np 2 \u00e2\u20ac\u201cenv MV2_USE_CUDA=1\u00c2\u00a0MV2_ENABLE_AFFINITY=0 \u00c2\u00a0.\/mpi app Cary MPICH2 Fine-grained Multi-threading Support Cray-MPICH now offers improved support for multi-threaded applications that perform MPI operations within threaded regions. Currently, this feature is available as a<\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[4],"tags":[],"class_list":["post-1525","post","type-post","status-publish","format-standard","hentry","category-cs"],"_links":{"self":[{"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/posts\/1525","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1525"}],"version-history":[{"count":1,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/posts\/1525\/revisions"}],"predecessor-version":[{"id":1800,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=\/wp\/v2\/posts\/1525\/revisions\/1800"}],"wp:attachment":[{"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1525"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1525"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/blog.jungwon.kim\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1525"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}
\nCray-MPICH now offers improved support for multi-threaded applications
\nthat perform MPI operations within threaded regions. Currently, this
\nfeature is available as a separate version of the Cray-MPICH library
\nthat is invoked by using a new compiler driver option, -craympich-mt.<\/p>\n
\nor higher, and supports both C\/C++ and Fortran in the PrgEnv-cray,
\nPrgEnv-intel, PrgEnv-gnu, and PrgEnv-pgi programming environments.<\/p>\n
\noperations. Applications that perform collective, one-sided, or MPI-
\nI\/O operations within threaded regions will not see any benefits.
\nSingle-threaded applications, or applications that do not currently
\nperform MPI operations within threaded regions, should continue to use
\nthe default Cray-MPICH library. This feature does not alter the
\nthread-safety behavior of the default Cray-MPICH library.<\/p>\n
\nfuture releases.<\/p>\n
\n-craympich-mt compiler driver option.<\/p>\n
\noptions are mutually exclusive. If both of these options are
\nspecified on a command line, the compiler will ignore the first
\none specified and use the second one.<\/p>\n
\ndisplay the version string and verify that the desired library has
\nbeen used. The default threading library displays this version string:<\/p>\n
\nglobal locks.<\/p>\n
\nwith per-object locks.<\/p>\n","protected":false},"excerpt":{"rendered":"