Merge branch 'open-mpi:main' into main

Author: amd-nithyavs

.github/workflows/compile-ze.yaml

@@ -0,0 +1,31 @@
+name: OneAPI ZE
+
+on: [pull_request]
+
+jobs:
+  compile-ze:
+    runs-on: ubuntu-22.04
+    steps:
+    - name: Install dependencies
+      run: |
+        sudo apt update
+        sudo apt install -y --no-install-recommends wget lsb-core software-properties-common gpg curl cmake git
+    - name: Build OneAPI ZE  
+      run: |
+        git clone https://github.com/oneapi-src/level-zero.git
+        cd level-zero
+        mkdir build
+        cd build
+        cmake ../ -DCMAKE_INSTALL_PREFIX=/opt/ze
+        sudo make -j install
+    - uses: actions/checkout@v3
+      with:
+            submodules: recursive
+    - name: Build Open MPI
+      run: |
+        ./autogen.pl
+        #
+        # we have to disable romio as its old ze stuff doesn't compile with supported ZE API
+        #
+        ./configure --prefix=${PWD}/install --disable-mpi-fortran --disable-io-romio --disable-oshmem --with-ze
+        make -j

.gitmodules

@@ -1,6 +1,6 @@
 [submodule "prrte"]
 	path = 3rd-party/prrte
-	url = ../../openpmix/prrte
+	url = ../../open-mpi/prrte
 	branch = master
 [submodule "openpmix"]
 	path = 3rd-party/openpmix

3rd-party/openpmix

@@ -11,6 +11,8 @@ dnl                         University of Stuttgart.  All rights reserved.
 dnl Copyright (c) 2004-2005 The Regents of the University of California.
 dnl                         All rights reserved.
 dnl Copyright (c) 2010-2014 Cisco Systems, Inc.  All rights reserved.
+dnl Copyright (c) 2024      Research Organization for Information Science
+dnl                         and Technology (RIST).  All rights reserved.
 dnl $COPYRIGHT$
 dnl
 dnl Additional copyrights may follow
@@ -35,6 +37,10 @@ SUBROUTINE binky(buf)
  REAL, DIMENSION(*), ASYNCHRONOUS :: buf
 END SUBROUTINE
 END INTERFACE
+CONTAINS
+SUBROUTINE wookie(buf)
+ REAL, DIMENSION(*), ASYNCHRONOUS :: buf
+END SUBROUTINE
 END MODULE asynch_mod]])],
              [AS_VAR_SET(asynchronous_var, yes)],
              [AS_VAR_SET(asynchronous_var, no)])

3rd-party/prrte

@@ -14,6 +14,8 @@ dnl Copyright (c) 2007      Los Alamos National Security, LLC.  All rights
 dnl                         reserved.
 dnl Copyright (c) 2007      Sun Microsystems, Inc.  All rights reserved.
 dnl Copyright (c) 2009-2015 Cisco Systems, Inc.  All rights reserved.
+dnl Copyright (c) 2024      Research Organization for Information Science
+dnl                         and Technology (RIST).  All rights reserved.
 dnl $COPYRIGHT$
 dnl
 dnl Additional copyrights may follow
@@ -82,6 +84,12 @@ AC_DEFUN([_OMPI_FORTRAN_CHECK_IGNORE_TKR], [
               [!GCC\$ ATTRIBUTES NO_ARG_CHECK ::], [type(*), dimension(*)],
               [!GCC\$ ATTRIBUTES NO_ARG_CHECK],
               [internal_ignore_tkr_happy=1], [internal_ignore_tkr_happy=0])])
+    # LLVM compilers
+    AS_IF([test $internal_ignore_tkr_happy -eq 0],
+          [OMPI_FORTRAN_CHECK_IGNORE_TKR_SUB(
+              [!DIR\$ IGNORE_TKR], [type(*)],
+              [!DIR\$ IGNORE_TKR],
+              [internal_ignore_tkr_happy=1], [internal_ignore_tkr_happy=0])])
     # Intel compilers
     AS_IF([test $internal_ignore_tkr_happy -eq 0],
           [OMPI_FORTRAN_CHECK_IGNORE_TKR_SUB(
@@ -133,6 +141,7 @@ AC_DEFUN([OMPI_FORTRAN_CHECK_IGNORE_TKR_SUB], [
     AC_MSG_CHECKING([for Fortran compiler support of $3])
     AC_COMPILE_IFELSE(AC_LANG_PROGRAM([],[[!
 ! Autoconf puts "program main" at the top
+ implicit none
 
   interface
      subroutine force_assumed_shape(a, count)
@@ -157,6 +166,7 @@ AC_DEFUN([OMPI_FORTRAN_CHECK_IGNORE_TKR_SUB], [
   complex, pointer, dimension(:,:) :: ptr
   target :: buffer3
   integer :: buffer4
+  integer :: a
   ptr => buffer3
 
 ! Set some known values (somewhat irrelevant for this test, but just be
@@ -189,8 +199,23 @@ AC_DEFUN([OMPI_FORTRAN_CHECK_IGNORE_TKR_SUB], [
     call foo(a, count)
   end subroutine force_assumed_shape
 
+  module check_ignore_tkr
+  interface
+     subroutine foobar(buffer, count)
+       $1 buffer
+       $2, intent(in) :: buffer
+       integer, intent(in) :: count
+     end subroutine foobar
+  end interface
+  end module
+
+  subroutine bar(var)
+    use check_ignore_tkr
+    implicit none
+    real, intent(inout) :: var(:, :, :)
+
+    call foobar(var(1,1,1), 1)
 ! Autoconf puts "end" after the last line
-  subroutine bogus
 ]]),
                     [msg=yes
                      ompi_fortran_ignore_tkr_predecl="$1"

config/ompi_fortran_check_asynchronous.m4

@@ -59,8 +59,8 @@ AC_ARG_WITH([cuda-libdir],
                             [Search for CUDA libraries in DIR])],
             [],
             [AS_IF([test -d "$with_cuda"],
-             [with_cuda_libdir=$(dirname $(find $with_cuda -name libcuda.so 2> /dev/null) 2> /dev/null)],
-             [with_cuda_libdir=$(dirname $(find /usr/local/cuda -name libcuda.so 2> /dev/null) 2> /dev/null)])
+             [with_cuda_libdir=$(dirname $(find -H $with_cuda -name libcuda.so 2> /dev/null) 2> /dev/null)],
+             [with_cuda_libdir=$(dirname $(find -H /usr/local/cuda -name libcuda.so 2> /dev/null) 2> /dev/null)])
             ])
 
 # Note that CUDA support is off by default.  To turn it on, the user has to

config/ompi_fortran_check_ignore_tkr.m4

@@ -120,7 +120,7 @@ information (adjust as necessary for your specific environment):
    # Fill in the options you want to pass to configure here
    options=""
    ./configure $options 2>&1  | tee $dir/config.out
-   tar -cf - `find . -name config.log` | tar -x -C $dir -
+   tar -cf - `find . -name config.log` | tar -x -C $dir
 
    # Build and install Open MPI
    make V=1 all 2>&1          | tee $dir/make.out

config/opal_check_cuda.m4

@@ -53,7 +53,7 @@ Fortran 2008 Syntax
 INPUT PARAMETERS
 ----------------
 * ``buf``: Initial address of send buffer (choice).
-* ``count``: Number of elements send (nonnegative integer).
+* ``count``: Number of elements in send buffer (nonnegative integer).
 * ``datatype``: Datatype of each send buffer element (handle).
 * ``dest``: Rank of destination (integer).
 * ``tag``: Message tag (integer).

docs/getting-help.rst

@@ -707,11 +707,6 @@ int ompi_comm_split_with_info( ompi_communicator_t* comm, int color, int key,
     /* Activate the communicator and init coll-component */
     rc = ompi_comm_activate (&newcomp, comm, NULL, NULL, NULL, false, mode);
 
-    /* MPI-4 §7.4.4 requires us to remove all unknown keys from the info object */
-    if (NULL != newcomp->super.s_info) {
-        opal_info_remove_unreferenced(newcomp->super.s_info);
-    }
-
  exit:
     free ( results );
     free ( sorted );
@@ -1028,9 +1023,6 @@ static int ompi_comm_split_type_core(ompi_communicator_t *comm,
         goto exit;
     }
 
-    /* MPI-4 §7.4.4 requires us to remove all unknown keys from the info object */
-    opal_info_remove_unreferenced(newcomp->super.s_info);
-
     /* TODO: there probably is better way to handle this case without throwing away the
      * intermediate communicator. */
     rc = ompi_comm_split (newcomp, local_split_type, key, newcomm, false);
@@ -1363,9 +1355,6 @@ int ompi_comm_dup_with_info ( ompi_communicator_t * comm, opal_info_t *info, omp
         return rc;
     }
 
-    /* MPI-4 §7.4.4 requires us to remove all unknown keys from the info object */
-    opal_info_remove_unreferenced(newcomp->super.s_info);
-
     *newcomm = newcomp;
     return MPI_SUCCESS;
 }
@@ -1522,8 +1511,6 @@ static int ompi_comm_idup_with_info_finish (ompi_comm_request_t *request)
 {
     ompi_comm_idup_with_info_context_t *context =
         (ompi_comm_idup_with_info_context_t *) request->context;
-    /* MPI-4 §7.4.4 requires us to remove all unknown keys from the info object */
-    opal_info_remove_unreferenced(context->newcomp->super.s_info);
 
     /* done */
     return MPI_SUCCESS;

docs/man-openmpi/man3/MPI_Send.3.rst

@@ -138,9 +138,6 @@ int ompi_file_open(struct ompi_communicator_t *comm, const char *filename,
         return ret;
     }
 
-    /* MPI-4 §14.2.8 requires us to remove all unknown keys from the info object */
-    opal_info_remove_unreferenced(file->super.s_info);
-
     /* All done */
 
     *fh = file;

ompi/communicator/comm.c

@@ -243,7 +243,7 @@ int ompi_mpiinfo_init_env(int argc, char *argv[], ompi_info_t *info)
 // related calls:
 
 int ompi_info_dup (ompi_info_t *info, ompi_info_t **newinfo) {
-    return opal_info_dup (&(info->super), (opal_info_t **)newinfo);
+    return opal_info_dup_public (&(info->super), (opal_info_t **)newinfo);
 }
 int ompi_info_set (ompi_info_t *info, const char *key, const char *value) {
     return opal_info_set (&(info->super), key, value);

ompi/file/file.c

@@ -2,15 +2,14 @@
 # Copyright (c) 2014      The University of Tennessee and The University
 #                         of Tennessee Research Foundation.  All rights
 #                         reserved.
-# Copyright (c) 2014      NVIDIA Corporation.  All rights reserved.
+# Copyright (c) 2014-2024 NVIDIA Corporation.  All rights reserved.
 # Copyright (c) 2017      IBM Corporation.  All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
 #
 # $HEADER$
 #
-dist_ompidata_DATA = help-mpi-coll-accelerator.txt
 
 sources = coll_accelerator_module.c coll_accelerator_reduce.c coll_accelerator_allreduce.c \
           coll_accelerator_reduce_scatter_block.c coll_accelerator_component.c \

ompi/info/info.c

@@ -2,7 +2,7 @@
  * Copyright (c) 2014      The University of Tennessee and The University
  *                         of Tennessee Research Foundation.  All rights
  *                         reserved.
- * Copyright (c) 2014-2015 NVIDIA Corporation.  All rights reserved.
+ * Copyright (c) 2014-2024 NVIDIA Corporation.  All rights reserved.
  * Copyright (c) 2024      Triad National Security, LLC. All rights reserved.
  * $COPYRIGHT$
  *
@@ -38,9 +38,6 @@ mca_coll_base_module_t
 *mca_coll_accelerator_comm_query(struct ompi_communicator_t *comm,
                           int *priority);
 
-int mca_coll_accelerator_module_enable(mca_coll_base_module_t *module,
-                                struct ompi_communicator_t *comm);
-
 int
 mca_coll_accelerator_allreduce(const void *sbuf, void *rbuf, int count,
                         struct ompi_datatype_t *dtype,