June 2015 - commits - mailman3.cels.anl.gov

[mpich] MPICH primary repository branch, master, updated. v3.2b3-134-g8b0b0a0
by noreply＠mpich.org 25 Jun '15

25 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 8b0b0a0f8371f39014b5b5a8e14db334f33bba46 (commit) via 2037eb5e9b228d8581dc02d33a5866d15c70e2f6 (commit) from 75adc6a13b918c57a79e38f7a04b7e6a2c607c60 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/8b0b0a0f8371f39014b5b5a8e14db334f… commit 8b0b0a0f8371f39014b5b5a8e14db334f33bba46 Author: Pavan Balaji <balaji(a)anl.gov> Date: Thu Jun 25 16:58:51 2015 -0500 Prepare for the 3.2b4 release. Signed-off-by: Halim Amer <aamer(a)anl.gov> diff --git a/CHANGES b/CHANGES index f96cde4..3ee36c0 100644 --- a/CHANGES +++ b/CHANGES @@ -31,7 +31,7 @@ A full list of changes is available at the following link: - http://git.mpich.org/mpich.git/shortlog/v3.1.3..v3.2b3 + http://git.mpich.org/mpich.git/shortlog/v3.1.3..v3.2b4 A full list of bugs that have been fixed is available at the following link: diff --git a/maint/version.m4 b/maint/version.m4 index 8e15768..6d35ed3 100644 --- a/maint/version.m4 +++ b/maint/version.m4 @@ -14,7 +14,7 @@ # changing this by playing with diversions, but then we would probably be # playing with autotools-fire. -m4_define([MPICH_VERSION_m4],[3.2b3])dnl +m4_define([MPICH_VERSION_m4],[3.2b4])dnl m4_define([MPICH_RELEASE_DATE_m4],[unreleased development copy])dnl # For libtool ABI versioning rules see: http://git.mpich.org/mpich.git/commitdiff/2037eb5e9b228d8581dc02d33a5866d15… commit 2037eb5e9b228d8581dc02d33a5866d15c70e2f6 Author: Pavan Balaji <balaji(a)anl.gov> Date: Thu Jun 25 16:58:38 2015 -0500 MX is deleted, not deprecated. Signed-off-by: Halim Amer <aamer(a)anl.gov> diff --git a/CHANGES b/CHANGES index 3802992..f96cde4 100644 --- a/CHANGES +++ b/CHANGES @@ -25,7 +25,7 @@ https://github.com/ofiwg/libfabric (thanks to Intel for code contribution) # The Myrinet MX network module, which had a life cyle from 1.1 till - 3.1.2, has now been deprecated. + 3.1.2, has now been deleted. # Several other minor bug fixes, memory leak fixes, and code cleanup. ----------------------------------------------------------------------- Summary of changes: CHANGES | 4 ++-- maint/version.m4 | 2 +- 2 files changed, 3 insertions(+), 3 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-132-g75adc6a
by noreply＠mpich.org 25 Jun '15

25 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 75adc6a13b918c57a79e38f7a04b7e6a2c607c60 (commit) from 5d421da641c2547831101fd0140f72ca369b95f7 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/75adc6a13b918c57a79e38f7a04b7e6a2… commit 75adc6a13b918c57a79e38f7a04b7e6a2c607c60 Author: Halim Amer <aamer(a)anl.gov> Date: Thu Jun 25 15:41:18 2015 -0500 Added a code cleanup script diff --git a/maint/code-cleanup.sh b/maint/code-cleanup.sh new file mode 100755 index 0000000..01c12ac --- /dev/null +++ b/maint/code-cleanup.sh @@ -0,0 +1,84 @@ +#! /bin/bash + +if test ! -z "`which gindent`" ; then + indent=gindent +else + indent=indent +fi + +indent_code() +{ + file=$1 + + $indent --k-and-r-style --line-length100 --else-endif-column1 --start-left-side-of-comments \ + --break-after-boolean-operator --dont-cuddle-else --dont-format-comments \ + --comment-indentation1 --indent-level4 --no-tabs \ + ${file} + rm -f ${file}~ + cp ${file} /tmp/${USER}.__tmp__ && \ + cat ${file} | sed -e 's/ *$//g' -e 's/( */(/g' -e 's/ *)/)/g' \ + -e 's/if(/if (/g' -e 's/while(/while (/g' -e 's/do{/do {/g' -e 's/}while/} while/g' > \ + /tmp/${USER}.__tmp__ && mv /tmp/${USER}.__tmp__ ${file} +} + +usage() +{ + echo "Usage: $1 [filename | --all] {--recursive} {--debug}" +} + +# Check usage +if [ -z "$1" ]; then + usage $0 + exit +fi + +# Make sure the parameters make sense +all=0 +recursive=0 +got_file=0 +debug= +ignore=0 +ignore_list="__I_WILL_NEVER_FIND_YOU__" +for arg in $@; do + if [ "$ignore" = "1" ] ; then + ignore_list="$ignore_list|$arg" + ignore=0 + continue; + fi + + if [ "$arg" = "--all" ]; then + all=1 + elif [ "$arg" = "--recursive" ]; then + recursive=1 + elif [ "$arg" = "--debug" ]; then + debug="echo" + elif [ "$arg" = "--ignore" ] ; then + ignore=1 + else + got_file=1 + fi +done +if [ "$recursive" = "1" -a "$all" = "0" ]; then + echo "--recursive cannot be used without --all" + usage $0 + exit +fi +if [ "$got_file" = "1" -a "$all" = "1" ]; then + echo "--all cannot be used in conjunction with a specific file" + usage $0 + exit +fi + +if [ "$recursive" = "1" ]; then + for i in `find . \! -type d | egrep '(\.c$|\.h$|\.c\.in$|\.h\.in$|\.cpp$|\.cpp.in$)' | \ + egrep -v "($ignore_list)"` ; do + ${debug} indent_code $i + done +elif [ "$all" = "1" ]; then + for i in `find . -maxdepth 1 \! -type d | egrep '(\.c$|\.h$|\.c\.in$|\.h\.in$|\.cpp$|\.cpp.in$)' | \ + egrep -v "($ignore_list)"` ; do + ${debug} indent_code $i + done +else + ${debug} indent_code $@ +fi ----------------------------------------------------------------------- Summary of changes: maint/code-cleanup.sh | 84 +++++++++++++++++++++++++++++++++++++++++++++++++ 1 files changed, 84 insertions(+), 0 deletions(-) create mode 100755 maint/code-cleanup.sh hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-131-g5d421da
by noreply＠mpich.org 24 Jun '15

24 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 5d421da641c2547831101fd0140f72ca369b95f7 (commit) via 45dfd81519787ef7559d5429bf92689c4cd58863 (commit) via 0b9a13e29e954a54087cf220bbed7aba4ee60d09 (commit) from 1dc5b1a50233fb7d22490daceaa36f596b49f357 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/5d421da641c2547831101fd0140f72ca3… commit 5d421da641c2547831101fd0140f72ca369b95f7 Author: Xin Zhao <xinzhao3(a)illinois.edu> Date: Fri Jun 19 20:52:20 2015 -0500 Bug-fix: correct SHM memory size in Nemesis Win_gather_info(). In the Nemesis implementation of Win_gather_info(), we allocate a memory region on SHM to store window information for other processes, so that all processes on the same node can share those information. However, previously the memory size was incorrectly set as O(node_comm_size), which should be O(comm_size). This patch fixed this bug. Signed-off-by: Min Si <msi(a)il.is.s.u-tokyo.ac.jp> Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/mpid/ch3/channels/nemesis/src/ch3_win_fns.c b/src/mpid/ch3/channels/nemesis/src/ch3_win_fns.c index 998aa95..eadd59d 100644 --- a/src/mpid/ch3/channels/nemesis/src/ch3_win_fns.c +++ b/src/mpid/ch3/channels/nemesis/src/ch3_win_fns.c @@ -350,7 +350,7 @@ static int MPIDI_CH3I_Win_gather_info(void *base, MPI_Aint size, int disp_unit, MPID_Comm * comm_ptr, MPID_Win ** win_ptr) { MPID_Comm *node_comm_ptr = NULL; - int node_rank, node_size; + int node_rank; int comm_rank, comm_size; MPI_Aint *tmp_buf = NULL; int i, k; @@ -371,10 +371,9 @@ static int MPIDI_CH3I_Win_gather_info(void *base, MPI_Aint size, int disp_unit, node_comm_ptr = (*win_ptr)->comm_ptr->node_comm; MPIU_Assert(node_comm_ptr != NULL); - node_size = node_comm_ptr->local_size; node_rank = node_comm_ptr->rank; - (*win_ptr)->info_shm_segment_len = node_size * sizeof(MPIDI_Win_basic_info_t); + (*win_ptr)->info_shm_segment_len = comm_size * sizeof(MPIDI_Win_basic_info_t); mpi_errno = MPIU_SHMW_Hnd_init(&(*win_ptr)->info_shm_segment_handle); if (mpi_errno != MPI_SUCCESS) http://git.mpich.org/mpich.git/commitdiff/45dfd81519787ef7559d5429bf92689c4… commit 45dfd81519787ef7559d5429bf92689c4cd58863 Author: Xin Zhao <xinzhao3(a)illinois.edu> Date: Tue Jun 23 21:41:09 2015 -0500 Bug-fix: correct local/remote completion in Win_flush_local. The original implementation in Win_flush_local counts number of total local completion and remote completion needed to wait, and then waiting for current local/remote completion count to reach those values. There is a bug that we should initialize the current count to zero in each while loop, otherwise the targets that are already completed will be count again and we failed to wait for some targets to be completed. This patch fixes this issue. Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/mpid/ch3/src/ch3u_rma_sync.c b/src/mpid/ch3/src/ch3u_rma_sync.c index 5fb34d8..c0f5689 100644 --- a/src/mpid/ch3/src/ch3u_rma_sync.c +++ b/src/mpid/ch3/src/ch3u_rma_sync.c @@ -1784,6 +1784,8 @@ int MPIDI_Win_flush_local_all(MPID_Win * win_ptr) /* wait for remote completion for those targets that disable flush_local, * and wait for local completion for other targets */ do { + local_completed_cnt = 0; + remote_completed_cnt = 0; for (i = 0; i < win_ptr->num_slots; i++) { curr_target = win_ptr->slots[i].target_list_head; while (curr_target != NULL) { http://git.mpich.org/mpich.git/commitdiff/0b9a13e29e954a54087cf220bbed7aba4… commit 0b9a13e29e954a54087cf220bbed7aba4ee60d09 Author: Xin Zhao <xinzhao3(a)illinois.edu> Date: Tue Jun 23 18:02:38 2015 -0500 Code format clean up. No reviewer. Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/mpid/ch3/src/mpid_rma.c b/src/mpid/ch3/src/mpid_rma.c index 0452fc3..0727994 100644 --- a/src/mpid/ch3/src/mpid_rma.c +++ b/src/mpid/ch3/src/mpid_rma.c @@ -395,7 +395,8 @@ static int win_init(MPI_Aint size, int disp_unit, int create_flavor, int model, if (MPIDI_RMA_Win_list == NULL) { mpi_errno = MPID_Progress_register_hook(MPIDI_CH3I_RMA_Make_progress_global); - if (mpi_errno) MPIU_ERR_POP(mpi_errno); + if (mpi_errno) + MPIU_ERR_POP(mpi_errno); } MPL_LL_APPEND(MPIDI_RMA_Win_list, MPIDI_RMA_Win_list_tail, win_elem); ----------------------------------------------------------------------- Summary of changes: src/mpid/ch3/channels/nemesis/src/ch3_win_fns.c | 5 ++--- src/mpid/ch3/src/ch3u_rma_sync.c | 2 ++ src/mpid/ch3/src/mpid_rma.c | 3 ++- 3 files changed, 6 insertions(+), 4 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-128-g1dc5b1a
by noreply＠mpich.org 24 Jun '15

24 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 1dc5b1a50233fb7d22490daceaa36f596b49f357 (commit) from e51b9c63dfa28d5dfd786d1f4ce21e51be0fc2fb (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/1dc5b1a50233fb7d22490daceaa36f596… commit 1dc5b1a50233fb7d22490daceaa36f596b49f357 Author: Charles J Archer <charles.j.archer(a)intel.com> Date: Tue Jun 23 21:41:43 2015 -0700 Progress registration fixup for pamid device Signed-off-by: Junchao Zhang <jczhang(a)mcs.anl.gov> diff --git a/src/mpid/common/sched/mpid_sched.c b/src/mpid/common/sched/mpid_sched.c index 52304e1..13c828e 100644 --- a/src/mpid/common/sched/mpid_sched.c +++ b/src/mpid/common/sched/mpid_sched.c @@ -944,7 +944,7 @@ int MPIDU_Sched_progress(int *made_progress) mpi_errno = MPIDU_Sched_progress_state(&all_schedules, made_progress); if (!mpi_errno && all_schedules.head == NULL) - MPIDI_CH3I_Progress_deregister_hook(MPIDU_Sched_progress); + MPID_Progress_deregister_hook(MPIDU_Sched_progress); return mpi_errno; } diff --git a/src/mpid/pamid/src/mpid_progress.c b/src/mpid/pamid/src/mpid_progress.c index 17191ff..1b6df50 100644 --- a/src/mpid/pamid/src/mpid_progress.c +++ b/src/mpid/pamid/src/mpid_progress.c @@ -26,7 +26,7 @@ typedef int (*progress_func_ptr_t) (int* made_progress); static progress_func_ptr_t progress_hooks[MAX_PROGRESS_HOOKS] = { NULL }; #undef FUNCNAME -#define FUNCNAME MPIDI_CH3I_Progress_register_hook +#define FUNCNAME MPIDI_Progress_register_hook #undef FCNAME #define FCNAME MPIDI_QUOTE(FUNCNAME) int MPIDI_Progress_register_hook(int (*progress_fn)(int*)) @@ -61,10 +61,10 @@ int MPIDI_Progress_register_hook(int (*progress_fn)(int*)) } #undef FUNCNAME -#define FUNCNAME MPIDI_CH3I_Progress_deregister_hook +#define FUNCNAME MPIDI_Progress_deregister_hook #undef FCNAME #define FCNAME MPIDI_QUOTE(FUNCNAME) -int MPIDI_CH3I_Progress_deregister_hook(int (*progress_fn)(int*)) +int MPIDI_Progress_deregister_hook(int (*progress_fn)(int*)) { int mpi_errno = MPI_SUCCESS; int i; ----------------------------------------------------------------------- Summary of changes: src/mpid/common/sched/mpid_sched.c | 2 +- src/mpid/pamid/src/mpid_progress.c | 6 +++--- 2 files changed, 4 insertions(+), 4 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-127-ge51b9c6
by noreply＠mpich.org 24 Jun '15

24 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via e51b9c63dfa28d5dfd786d1f4ce21e51be0fc2fb (commit) from df133f8eaf82f448e8d5fb53d0a8ec77d8484c85 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/e51b9c63dfa28d5dfd786d1f4ce21e51b… commit e51b9c63dfa28d5dfd786d1f4ce21e51be0fc2fb Author: Rob Latham <robl(a)mcs.anl.gov> Date: Tue Jun 23 19:36:51 2015 -0500 clean up 62829ab8: generate testlist properly No Reviewer diff --git a/test/mpi/coll/Makefile.am b/test/mpi/coll/Makefile.am index 6c7a2ed..0c1c5e1 100644 --- a/test/mpi/coll/Makefile.am +++ b/test/mpi/coll/Makefile.am @@ -7,7 +7,7 @@ include $(top_srcdir)/Makefile.mtest -EXTRA_DIST = testlist +EXTRA_DIST = testlist.in ## for all programs that are just built from the single corresponding source ## file, we don't need per-target _SOURCES rules, automake will infer them diff --git a/test/mpi/coll/testlist.in b/test/mpi/coll/testlist.in index 72da083..1c465b6 100644 --- a/test/mpi/coll/testlist.in +++ b/test/mpi/coll/testlist.in @@ -78,7 +78,7 @@ exscan 10 exscan2 5 gather 4 gather2 4 -@largetest@ gather_big 8 +@largetest@gather_big 8 scattern 4 scatter2 4 scatter3 4 diff --git a/test/mpi/configure.ac b/test/mpi/configure.ac index 7715ab9..9408bc1 100644 --- a/test/mpi/configure.ac +++ b/test/mpi/configure.ac @@ -1516,6 +1516,7 @@ AC_OUTPUT(maint/testmerge \ attr/Makefile \ util/Makefile \ coll/Makefile \ + coll/testlist \ comm/Makefile \ datatype/Makefile \ datatype/testlist \ ----------------------------------------------------------------------- Summary of changes: test/mpi/coll/Makefile.am | 2 +- test/mpi/coll/testlist.in | 2 +- test/mpi/configure.ac | 1 + 3 files changed, 3 insertions(+), 2 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-126-gdf133f8
by noreply＠mpich.org 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via df133f8eaf82f448e8d5fb53d0a8ec77d8484c85 (commit) from a4aa067596105b9b21d80c672eb00820e4c84cb5 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/df133f8eaf82f448e8d5fb53d0a8ec77d… commit df133f8eaf82f448e8d5fb53d0a8ec77d8484c85 Author: Junchao Zhang <jczhang(a)mcs.anl.gov> Date: Mon Jun 15 14:50:27 2015 -0500 Fix typos in an F08 test No reviewer diff --git a/test/mpi/f08/attr/attrmpi1f08.f90 b/test/mpi/f08/attr/attrmpi1f08.f90 index 242ee1b..fe5e2e9 100644 --- a/test/mpi/f08/attr/attrmpi1f08.f90 +++ b/test/mpi/f08/attr/attrmpi1f08.f90 @@ -5,7 +5,8 @@ ! See COPYRIGHT in top-level directory. ! program main - use mpi + use mpi_f08 + implicit none integer value, wsize, wrank, extra, mykey integer rvalue, svalue TYPE(MPI_Comm) ncomm @@ -19,7 +20,7 @@ ! ! Simple attribute put and get ! - call mpi_keyval_create( MPI_NULL_COPY_FN, MPI_NULL_DELETE_FN, & + call mpi_keyval_create( MPI_COMM_NULL_COPY_FN, MPI_COMM_NULL_DELETE_FN, & & mykey, extra,ierr ) call mpi_attr_get( MPI_COMM_WORLD, mykey, value, flag, ierr ) if (flag) then ----------------------------------------------------------------------- Summary of changes: test/mpi/f08/attr/attrmpi1f08.f90 | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-125-ga4aa067
by noreply＠mpich.org 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via a4aa067596105b9b21d80c672eb00820e4c84cb5 (commit) from 62829ab8a102daf287178e3fa4844cb19a2410a9 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/a4aa067596105b9b21d80c672eb00820e… commit a4aa067596105b9b21d80c672eb00820e4c84cb5 Author: Rob Latham <robl(a)mcs.anl.gov> Date: Tue Jun 23 09:31:40 2015 -0500 Increase number of finalizers Lisandro Dalcin <dalcinl(a)gmail.com> reports that mpi4py's test suite invokes MPIR_Add_finalize() 33 times. It's been 6.5 years since we doubled it, so bump it up once again. Closes: #2272 Signed-off-by: Junchao Zhang <jczhang(a)mcs.anl.gov> diff --git a/src/mpi/init/finalize.c b/src/mpi/init/finalize.c index f71668b..3aeeca7 100644 --- a/src/mpi/init/finalize.c +++ b/src/mpi/init/finalize.c @@ -65,7 +65,7 @@ typedef struct Finalize_func_t { } Finalize_func_t; /* When full debugging is enabled, each MPI handle type has a finalize handler installed to detect unfreed handles. */ -#define MAX_FINALIZE_FUNC 32 +#define MAX_FINALIZE_FUNC 64 static Finalize_func_t fstack[MAX_FINALIZE_FUNC]; static int fstack_sp = 0; static int fstack_max_priority = 0; @@ -76,7 +76,8 @@ void MPIR_Add_finalize( int (*f)( void * ), void *extra_data, int priority ) if (fstack_sp >= MAX_FINALIZE_FUNC) { /* This is a little tricky. We may want to check the state of MPIR_Process.mpich_state to decide how to signal the error */ - (void)MPIU_Internal_error_printf( "overflow in finalize stack!\n" ); + (void)MPIU_Internal_error_printf( "overflow in finalize stack! " + "Is MAX_FINALIZE_FUNC too small?\n" ); if (OPA_load_int(&MPIR_Process.mpich_state) == MPICH_IN_INIT || OPA_load_int(&MPIR_Process.mpich_state) == MPICH_POST_INIT) { ----------------------------------------------------------------------- Summary of changes: src/mpi/init/finalize.c | 5 +++-- 1 files changed, 3 insertions(+), 2 deletions(-) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-124-g62829ab
by noreply＠mpich.org 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 62829ab8a102daf287178e3fa4844cb19a2410a9 (commit) from 7ddceb6969857851d06033a6f1a9654b5f83acc5 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/62829ab8a102daf287178e3fa4844cb19… commit 62829ab8a102daf287178e3fa4844cb19a2410a9 Author: Rob Latham <robl(a)mcs.anl.gov> Date: Tue Jun 23 14:07:03 2015 -0500 mark gather_big as a 'largetest' disable this memory-intensive test on 32 bit platforms No Reviewer diff --git a/test/mpi/coll/testlist b/test/mpi/coll/testlist.in similarity index 99% rename from test/mpi/coll/testlist rename to test/mpi/coll/testlist.in index 057b96f..72da083 100644 --- a/test/mpi/coll/testlist +++ b/test/mpi/coll/testlist.in @@ -78,7 +78,7 @@ exscan 10 exscan2 5 gather 4 gather2 4 -gather_big 8 +@largetest@ gather_big 8 scattern 4 scatter2 4 scatter3 4 ----------------------------------------------------------------------- Summary of changes: test/mpi/coll/{testlist => testlist.in} | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-) rename test/mpi/coll/{testlist => testlist.in} (99%) hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, signed/rma-test-so, deleted. v3.2b3-114-g14099b3
by noreply＠mpich.org 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, signed/rma-test-so has been deleted was 14099b3e7981be79fc920b54ed0a8f2d60cc9a7b ----------------------------------------------------------------------- 14099b3e7981be79fc920b54ed0a8f2d60cc9a7b Added tests to stream ACC-like RMA ops ----------------------------------------------------------------------- hooks/post-receive -- MPICH primary repository

1 0

[mpich] MPICH primary repository branch, master, updated. v3.2b3-123-g7ddceb6
by noreply＠mpich.org 23 Jun '15

23 Jun '15

This is an automated email from the git hooks/post-receive script. It was generated because a ref change was pushed to the repository containing the project "MPICH primary repository". The branch, master has been updated via 7ddceb6969857851d06033a6f1a9654b5f83acc5 (commit) via 036d91a9927ccf5bb38d9ecc21882684319c46cf (commit) via a500a7dae4b5d10c42904751f7f5a45b7fb249cd (commit) from bac25133c02822ef28e5011a71ad713966c5f0c5 (commit) Those revisions listed above that are new to this repository have not appeared on any other notification email; so we list those revisions in full, below. - Log ----------------------------------------------------------------- http://git.mpich.org/mpich.git/commitdiff/7ddceb6969857851d06033a6f1a9654b5… commit 7ddceb6969857851d06033a6f1a9654b5f83acc5 Author: Rob Latham <robl(a)mcs.anl.gov> Date: Wed Jun 10 15:13:45 2015 -0500 adjust flattened processing for repeated resize resize and struct are the two type constructors that can set the LB and UB markers on a type. Struct, due to MPI-1 ideas, is a strange beast (they adjust only if they are lower/higher than the old ones (!) ) but for resized it's clear that the markers shift. Closes #2088 Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/mpi/romio/adio/common/flatten.c b/src/mpi/romio/adio/common/flatten.c index beae9e5..1a2bf1a 100644 --- a/src/mpi/romio/adio/common/flatten.c +++ b/src/mpi/romio/adio/common/flatten.c @@ -58,6 +58,7 @@ void ADIOI_Flatten_datatype(MPI_Datatype datatype) flat->next = NULL; flat->blocklens = NULL; flat->indices = NULL; + flat->lb_idx = flat->ub_idx = -1; flat->count = ADIOI_Count_contiguous_blocks(datatype, &curr_index); #ifdef FLATTEN_DEBUG @@ -108,6 +109,7 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, ADIO_Offset st_offset, MPI_Count *curr_index) { int i, k, m, n, basic_num, nonzeroth, is_hindexed_block=0; + int lb_updated=0; int combiner, old_combiner, old_is_contig; int nints, nadds, ntypes, old_nints, old_nadds, old_ntypes; /* By using ADIO_Offset we preserve +/- sign and @@ -736,18 +738,31 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, /* handle the Lb */ j = *curr_index; - flat->indices[j] = st_offset + adds[0]; - /* this zero-length blocklens[] element, unlike eleswhere in the - * flattening code, is correct and is used to indicate a lower bound - * marker */ - flat->blocklens[j] = 0; - flat->lb_idx = *curr_index; - - #ifdef FLATTEN_DEBUG - DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",0,adds[0],j, flat->indices[j], j, flat->blocklens[j]); - #endif + /* when we process resized types, we (recursively) process the lower + * bound, the type being resized, then the upper bound. In the + * resized-of-resized case, we might find ourselves updating the upper + * bound based on the inner type, but the lower bound based on the + * upper type. check both lb and ub to prevent mixing updates */ + if (flat->lb_idx == -1 && flat->ub_idx == -1) { + flat->indices[j] = st_offset + adds[0]; + /* this zero-length blocklens[] element, unlike eleswhere in the + * flattening code, is correct and is used to indicate a lower bound + * marker */ + flat->blocklens[j] = 0; + flat->lb_idx = *curr_index; + lb_updated=1; + + #ifdef FLATTEN_DEBUG + DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",0,adds[0],j, flat->indices[j], j, flat->blocklens[j]); + #endif - (*curr_index)++; + (*curr_index)++; + } else { + /* skipped over this chunk because something else higher-up in the + * type construction set this for us already */ + flat->count--; + st_offset -= adds[0]; + } /* handle the datatype */ @@ -773,12 +788,20 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, } /* take care of the extent as a UB */ - j = *curr_index; - flat->indices[j] = st_offset + adds[0] + adds[1]; - /* again, zero-element ok: an upper-bound marker explicitly set by the - * constructor of this resized type */ - flat->blocklens[j] = 0; - flat->lb_idx = *curr_index; + /* see note above about mixing updates for why we check lb and ub */ + if ((flat->lb_idx == -1 && flat->ub_idx == -1) || lb_updated) { + j = *curr_index; + flat->indices[j] = st_offset + adds[0] + adds[1]; + /* again, zero-element ok: an upper-bound marker explicitly set by the + * constructor of this resized type */ + flat->blocklens[j] = 0; + flat->ub_idx = *curr_index; + } else { + /* skipped over this chunk because something else higher-up in the + * type construction set this for us already */ + flat->count--; + (*curr_index)--; + } #ifdef FLATTEN_DEBUG DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",1,adds[1],j, flat->indices[j], j, flat->blocklens[j]); http://git.mpich.org/mpich.git/commitdiff/036d91a9927ccf5bb38d9ecc218826843… commit 036d91a9927ccf5bb38d9ecc21882684319c46cf Author: Rob Latham <robl(a)mcs.anl.gov> Date: Thu Jun 4 16:43:55 2015 -0500 record lb and ub index in flattened in deep types we might want to update the lb and ub, not simply append/prepend two tuples to the flattened representation. Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/mpi/romio/adio/common/flatten.c b/src/mpi/romio/adio/common/flatten.c index 86c2875..beae9e5 100644 --- a/src/mpi/romio/adio/common/flatten.c +++ b/src/mpi/romio/adio/common/flatten.c @@ -693,6 +693,10 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, flat->indices[j] = st_offset + adds[n]; MPI_Type_size_x(types[n], &old_size); flat->blocklens[j] = blocklength * old_size; + if (types[n] == MPI_LB) + flat->lb_idx = j; + if (types[n] == MPI_UB) + flat->ub_idx = j; #ifdef FLATTEN_DEBUG DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",n,adds[n],j, flat->indices[j], j, flat->blocklens[j]); #endif @@ -737,6 +741,7 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, * flattening code, is correct and is used to indicate a lower bound * marker */ flat->blocklens[j] = 0; + flat->lb_idx = *curr_index; #ifdef FLATTEN_DEBUG DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",0,adds[0],j, flat->indices[j], j, flat->blocklens[j]); @@ -773,6 +778,7 @@ void ADIOI_Flatten(MPI_Datatype datatype, ADIOI_Flatlist_node *flat, /* again, zero-element ok: an upper-bound marker explicitly set by the * constructor of this resized type */ flat->blocklens[j] = 0; + flat->lb_idx = *curr_index; #ifdef FLATTEN_DEBUG DBG_FPRINTF(stderr,"ADIOI_Flatten:: simple adds[%#X] "MPI_AINT_FMT_HEX_SPEC", flat->indices[%#llX] %#llX, flat->blocklens[%#llX] %#llX\n",1,adds[1],j, flat->indices[j], j, flat->blocklens[j]); diff --git a/src/mpi/romio/adio/include/adioi.h b/src/mpi/romio/adio/include/adioi.h index d2dbc94..555af6d 100644 --- a/src/mpi/romio/adio/include/adioi.h +++ b/src/mpi/romio/adio/include/adioi.h @@ -114,6 +114,16 @@ typedef struct ADIOI_Fl_node { MPI_Count count; /* no. of contiguous blocks */ ADIO_Offset *blocklens; /* array of contiguous block lengths (bytes)*/ ADIO_Offset *indices; /* array of byte offsets of each block */ + /* the type processing code in ROMIO loops through the flattened + * representation to tile file views. so, we cannot simply indicate a + * lower bound and upper bound with entries here -- those are instead + * indicated by offset-length pairs with zero length. In order to deal + * with repeatedly resetting the LB and UB though (as in resized of + * resized, or struct of resized, perhaps?), indicate where in the + * tuple-stream the lb and ub parts are kept + * (-1 indicates "not explicitly set") */ + ADIO_Offset lb_idx; + ADIO_Offset ub_idx; struct ADIOI_Fl_node *next; /* pointer to next node */ } ADIOI_Flatlist_node; http://git.mpich.org/mpich.git/commitdiff/a500a7dae4b5d10c42904751f7f5a45b7… commit a500a7dae4b5d10c42904751f7f5a45b7fb249cd Author: Rob Latham <robl(a)mcs.anl.gov> Date: Wed May 20 15:06:36 2015 -0500 hoist ROMIO init into MPI_Init some libraries like HDF5 want to register their cleanup routines into finalize. these cleanup routines use MPI-IO, so they need to fire before ROMIO cleans up. Signed-off-by: Pavan Balaji <balaji(a)anl.gov> diff --git a/src/include/mpiimpl.h b/src/include/mpiimpl.h index a6cf10d..42979d6 100644 --- a/src/include/mpiimpl.h +++ b/src/include/mpiimpl.h @@ -4341,6 +4341,7 @@ int MPIR_Type_size_x_impl(MPI_Datatype datatype, MPI_Count *size); /* random initializers */ int MPIR_Group_init(void); int MPIR_Comm_init(MPID_Comm *); +int MPIR_MPIOInit(int *error); /* Collective functions cannot be called from multiple threads. These diff --git a/src/mpi/init/initthread.c b/src/mpi/init/initthread.c index 6be0c7b..1e876b3 100644 --- a/src/mpi/init/initthread.c +++ b/src/mpi/init/initthread.c @@ -560,6 +560,10 @@ int MPIR_Init_thread(int * argc, char ***argv, int required, int * provided) if (mpi_errno == MPI_SUCCESS) mpi_errno = MPID_InitCompleted(); +#ifdef HAVE_ROMIO + MPIR_MPIOInit(&mpi_errno); +#endif + fn_exit: MPIU_THREAD_CS_EXIT(INIT,required); /* Make fields of MPIR_Process global visible and set mpich_state diff --git a/src/mpi/romio/mpi-io/mpir-mpioinit.c b/src/mpi/romio/mpi-io/mpir-mpioinit.c index be78b45..9a5512a 100644 --- a/src/mpi/romio/mpi-io/mpir-mpioinit.c +++ b/src/mpi/romio/mpi-io/mpir-mpioinit.c @@ -19,23 +19,8 @@ extern int ADIO_Init_keyval; void MPIR_MPIOInit(int * error_code) { - int flag; - char myname[] = "MPIR_MPIOInit"; - /* first check if ADIO has been initialized. If not, initialize it */ if (ADIO_Init_keyval == MPI_KEYVAL_INVALID) { - MPI_Initialized(&flag); - - /* --BEGIN ERROR HANDLING-- */ - if (!flag) { - *error_code = MPIO_Err_create_code(MPI_SUCCESS, - MPIR_ERR_RECOVERABLE, myname, __LINE__, - MPI_ERR_OTHER, "**initialized", 0); - *error_code = MPIO_Err_return_file(MPI_FILE_NULL, *error_code); - return; - } - /* --END ERROR HANDLING-- */ - MPI_Keyval_create(MPI_NULL_COPY_FN, ADIOI_End_call, &ADIO_Init_keyval, (void *) 0); ----------------------------------------------------------------------- Summary of changes: src/include/mpiimpl.h | 1 + src/mpi/init/initthread.c | 4 ++ src/mpi/romio/adio/common/flatten.c | 59 +++++++++++++++++++++++++-------- src/mpi/romio/adio/include/adioi.h | 10 ++++++ src/mpi/romio/mpi-io/mpir-mpioinit.c | 15 -------- 5 files changed, 59 insertions(+), 30 deletions(-) hooks/post-receive -- MPICH primary repository

1 0