Proposed extensions for Open MPI:
- If MPI_INITLIZED is invoked and MPI is only partially initialized,
wait until MPI is fully initialized before returning.
- If MPI_FINALIZED is invoked and MPI is only partially finalized,
wait until MPI is fully finalized before returning.
- If the ompi_mpix_allow_multi_init MCA param is true, allow MPI_INIT
and MPI_INIT_THREAD to be invoked multiple times without error (MPI
will be safely initialized only the first time it is invoked).
Once a FIN control message is appended to the pending list,
the ob1 PML attempts to send the FIN again in the `mca_pml_ob1_process_pending_packets` function.
But if the PML failed to sent the FIN again, the `mca_pml_ob1_send_fin`
function creates a new `mca_pml_ob1_pckt_pending_t` object and the
old object is not retured to the free list.
Add ompi_mpi_dynamics_disable() function to disable MPI dynamic
process functionality (i.e., such that if MPI_COMM_SPAWN/etc. are
invoked, you'll get a show_help error explaining that MPI dynamic
process functionality is disabled in this environment -- instead of a
potentially-cryptic network or hardware error).
Fixes#984
This commit fixes an issue identified by @rolfv. The local peer was
not being correctly initialized when running with a single process on
a node.
This fixesopen-mpi/ompi#1010
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds implementations of gather and igather using
Portals4 triggered operations. The default algorithm is linear,
but binomial can be selected using an MCA parameter -
coll_portals4_use_binomial_gather_algorithm.
when profiling is built.
This prevents oshmem subroutines from being wrapped twice by third
party tools (e.g. once in oshmem and once in MPI)
see discussion starting at http://www.open-mpi.org/community/lists/devel/2015/08/17842.php
Thanks to Bert Wesarg for bringing this to our attention
This patch fixes the issues identified by @ggouaillardet in the IBM tests (collectives and topologies). It also improves the memory usage of OMPI, as a communicator without collective communications will never allocate the array of requests needed to coordinate the basic collective algorithms. This ticket replaced #790.
Fix code paths that didn't convert the MPI datatype to the
corresponding Portals4 datatype.
Thanks to Nicolas Chevalier (@shawone) for finding this bug and
submitting a patch.
in the base).
Correctly deal with persistent requests (they must be always freed when
they are stored in the request array associated with the communicator).
Always use MPI_STATUS_IGNORE for single request waiting functions.
The Portals4 get_peer family incorrectly cast the ompi_proc_t to
ptl_process_t and returned that as the peer. The ptl_process_t is
actually found in the endpoint array. This commit fixes the
Portals4 get_peer family to return the dereferenced endpoint
pointer.
This commit frees any outstanding windows at ompi_win_finalize. If
ompi_debug_show_handle_leaks is set a warning message is printed out
indicating that a window is still allocated.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
FCA barrier may not complete if FCA progress is not called periodically.
PMI/PMI2 API that can be used in rte barrier has no provision for calling
external progress function.
So it is possible that during finalize some ranks will be stuck
in fca barrier while others are in PMI barrier.
Fixes CID 72320: Explicit NULL dereferenced
On error it is possible that the blocklen_per_process array is
NULL. Change the NULL check before the free to check for non-NULL on
the array not the array element. Also clean up allocation of this
array to use calloc instead of malloc + setting each element to NULL.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fixes CIDs 72300, 72344, 1196764-1196768, 72300: Resource leaks
Mulitple allocated arrays are going out of scope at the end of
mca_fcoll_two_phase_file_write_all. Free these arrays. Also removed
the extraneous NULL checks since free (NULL) is safe in C.
Change returns to goto exit where the allocated resources are freed.
Fixes CIDs 72285-72292, 72297, 72298: Resource leaks
Change all appropriate return statements to goto exit to ensure that
all resources are freed. Also removed the NULL checks since free
(NULL) is safe in C.
Fixes CIDs 72295, 72296: Resource leaks
Moved free of requests and recv_types to after exit label. This will
ensure these are freed on error.
Also added a loop and statement to free send_buf which is going out of
scope at the end of the function.
Fixes CIDs 72336-72240, 735197, 735198: Resource leaks
Moved the exit label before to before the resources are released and
changed all appropriate return statements to goto exit. Also removed
extraneous NULL checks because free (NULL) is safe in C.
Fixes CIDs 72341, 72343, 1196805-1196809: Resource leaks
Free all resources after exit label and change return statements to
goto exit to ensure all resources are freed on error.
Fixes CID 1269973: Unused value
Check return code of ompi_request_wait_all. If it fails jump to the
exit.
Fixes CID 714119: Dereference before NULL check
Wrong value checked in conditional.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fixes CID 1323841: Logically dead code
Wrong value in conditional. Should be newcomp not newcomm.
Fixes CID 1269762: Explicit null dereference
ompi_group_incl could return an error and not set local_group. Add a
check to ensure ompi_group_incl succeeded before incrementing the proc
count.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fix CID 1315271: Constant expression result
The intent of this conditional is to not produce a peruse event for
probe or mprobe requests. Coverity is correct that the expression is
always true. Changed the || to && to fix. Also moved the conditional
within an OMPI_WANT_PERUSE to ensure the conditional is not evaluated
if peruse is disabled.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fixed CID 1269712, 1269709, 1269706, 1269703, 1269694: Logically dead code
Remove extra NULL check as OMPI_OSC_PT2PT_REQUEST_ALLOC can never set the
request to NULL.
Fixes CID 1269668: Unchecked return value
False positive. Add (void) to indicate we do not care about the return code
from opal_hash_table_get_uint32.
Fixes CID 1324726: Free of address-of expression
Do not free lock if it was not allocated.
Fixes CID 1269658: Free of address-of expression
Never will happen but because op is always a built-in op there is no
reason to retain/release it anyway.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
In the default mode of operation, the Portals4 components support
dynamic add_procs().
The Portals4 components have two alternate modes (flow control and
logical-to-physical) that require knowledge of all procs at startup.
In these modes, mtl-portals4 sets the MCA_MTL_BASE_FLAG_REQUIRE_WORLD
flag and btl-portals4 sets the MCA_BTL_FLAGS_SINGLE_ADD_PROCS flag
to tell the PML that we need all the procs in one add_procs() call.
This commit adds support to the pml, mtl, and btl frameworks for
components to indicate at runtime that they do not support the new
dynamic add_procs behavior. At the high end the lack of dynamic
add_procs support is signalled by the pml using the new pml_flags
member to the pml module structure. If the
MCA_PML_BASE_FLAG_REQUIRE_WORLD flag is set MPI_Init will generate the
ompi_proc_t array passed to add_proc from ompi_proc_world () instead
of ompi_proc_get_allocated ().
Both cm and ob1 have been updated to detect if the underlying mtl and
btl components support dynamic add_procs.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds two new functions:
- ompi_proc_get_allocated - Returns all procs in the current job that
have already been allocated. This is used in init/finalize to
determine which procs to pass to add_procs/del_procs.
- ompi_proc_world_size - returns the number of processes in
MPI_COMM_WORLD. This may be removed in favor of callers just
looking at ompi_process_info.
The behavior of ompi_proc_world has been restored to return
ompi_proc_t's for all processes in the current job. The use of this
function is discouraged.
Code that was using ompi_proc_world() has been updated to make use of
the new functions to avoid the memory overhead of ompi_comm_world ().
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fix CID 1196812: Resource Leak
dsts array was leaked on error.
Fix CID 710565: Copy-paste error
The line in question (nbc:513) is indeed a copy-paste error.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
ROMIO configure looks for lstat in wrong header
The ROMIO configure script checks for a declaration of lstat in
unistd.h, but, at least on the Linux machines I checked, lstat is in
sys/stat.h. (The detection failure led to a linker error when building
ROMIO as part of OpenMPI on one of my admittedly strangely configured
machines, somehow.) It appears from the man page that either location
is possible, so check both.
(cherry picked from mpich/mpich@7b8bd055df)
Signed-off-by: Rob Latham <robl@mcs.anl.gov>
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The add_procs change made some assumptions in the bml/r2 add_procs
wrong. This lead to del_procs never being called. I removed the logic
that checks the ompi_proc_t reference count and removed an unnecessary
allocation. The allocation only makes sense if we pass more than a
single proc at a time to the btl del_procs.
This commit also ensures that the btl del_procs is called if the
endpoint is in the btl_rdma array but not the btl_send array.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The osc/sm component was using a simple counter to determine if all
expected posts had arrived to start a PSCW access epoch. This is
incorrect as a post may arrive from a peer that isn't part of the
current start group. There are many ways this could have been fixed.
This commit adds an n^2 bitmap. When a process posts it sets a bit in
the bitmap associated with the access rank to indicate the post is
complete. The access rank checks for and clears the bits associated
with all the processes in the start group.
The bitmap requires comm_size ^ 2 bits of space. This should be
managable as most nodes have relatively small numbers of processes. If
this changes another algorigthm can be implemented.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Fix CID 1324733: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324734: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324735: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324736: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324737: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324751: Memory - illegal accesses (USE_AFTER_FREE)
Fix CID 1324750: (USE_AFTER_FREE)
Fix CID 1324749: Memory - corruptions (USE_AFTER_FREE)
Fix CID 1324748: Memory - illegal accesses (USE_AFTER_FREE)
Fix CID 1324747: (USE_AFTER_FREE)
Fix CID 1324746: Memory - corruptions (USE_AFTER_FREE)
Add missing return on an error path.
Fix CID 1324745: Code maintainability issues (UNUSED_VALUE)
Ignore return code from barrier. It was not being used anyway.
Fix CID 1324738: Null pointer dereferences (FORWARD_NULL)
Fix CID 1324741: Null pointer dereferences (REVERSE_INULL)
module->selected_btl can not be NULL in osc/rdma during normal
operation. Removed the unnecessary NULL check.
Fix CID 1324752: Memory - illegal accesses (USE_AFTER_FREE)
Move ompi_osc_pt2pt_module_lock_remove to before the lock is freed.
Fix CID 1324744: Uninitialized variables (UNINIT)
Fix CID 1324743: Uninitialized variables (UNINIT)
This array is not used unitialized but there is no reason not to use
calloc here to silence the warning.
The following CID is a false positive: 1324742. I will mark it such in
coverity.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds man pages for the MPI_Win_allocate and MPI_Win_allocated_shared
MPI-3 functions. The man page for MPI_Win_create has also been updated to
indicate support for the same_size and same_disp_unit info keys
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Most functionality of oshmem_proc duplicates ompi_proc. In addition
to that, Current logic does not allow to do oshmem initialization
w/o ompi startup.
So this refactoring allows to avoid code duplication, decrease used
memory and make oshmem support easier.
Now oshmem_proc is transparent ompi_proc structure, that can be
extended by oshmem specific data.
Signed-off-by: Igor Ivanov <Igor.Ivanov@itseez.com>
This commit adds support for performing one-sided operations over
supported hardware (currently Infiniband and Cray Gemini/Aries). This
component is still undergoing active development.
Current features:
- Use network atomic operations (fadd, cswap) for implementing
locking and PSCW synchronization.
- Aggregate small contiguous puts.
- Reduced memory footprint by storing window data (pointer, keys,
etc) at the lowest rank on each node. The data is fetched as each
process needs to communicate with a new peer. This is a trade-off
between the performance of the first operation on a peer and the
memory utilization of a window.
TODO:
- Add support for the accumulate_ops info key. If it is known that
the same op or same op/no op is used it may be possible to use
hardware atomics for fetch-and-op and compare-and-swap.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit updates osc/pt2pt to allocate peer object as they are
needed rather than all at once. Additionally, to help improve the
memory footprint a new synchronization structure has been added.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The behavior of ompi_comm_set was changed to get the remote size from
the remote group. This broke how ompi_comm_[i]dup were using
ompi_comm_set. In order to adapt to the new behavior these functions
now pass NULL for the remote group if the communicator is not an
inter-communicator.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit updates the behavior of ompi_comm_set to explicitly take
either local/remote group(s) OR local/remote array(s). If array(s) are
in use the sizes will be taken from the appropriate group(s).
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit includes two changes. First, the locality code has been
factored out to improve readability and maintainability. Second,
instead of looking up each proc using ompi_group_peer_lookup the code
now uses ompi_group_peer_lookup_existing. The code falls back on modex
if a proc doesn't exist. This will prevent MPI_Comm_split_type from
allocating ompi_proc_t's for every process in the job.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The assumption that the high bit is not in use in pointers on any of our
supported platforms was incorrect. A better assumption is that all
ompi_proc_t pointers will be at least 2-byte aligned. This allows us
to use the low bit. To do this we drop the highest bit of the
opal_process_name_t jobid (hope this is ok) and use the low bit to
indicate the proc is really a sentinel.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
This commit removes the use of ompi_group_peer_lookup in the
ompi_dpm_mark_dyncomm function. The function now uses
ompi_group_get_proc_name which does not allocate an ompi_proc_t if one
does not already exist.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Add an accessor for the proc_endpoints[OMPI_PROC_ENDPOINT_TAG_MTL]
member of the ompi_proc_t structure. This accessort calls add_procs
with the ompi_proc_t if the member is NULL.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
Add an accessor for the proc_endpoints[OMPI_PROC_ENDPOINT_TAG_MTL]
member of the ompi_proc_t structure. This accessort calls add_procs
with the ompi_proc_t if the member is NULL. Tested on an infinipath
system with InfiniPath_QLE7340 HCAs.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
Updated the union/difference code to remove an extra n^2 translation
of ranks. This comes at the cost of extra memory but greatly
simplifies the code.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
This commit modifies the ompi_group_t union/difference code to compare/copy the
raw group values. This will either be a ompi_proc_t or a sentinel value. This
commit also adds helper functions to convert between opal process names and
sentinel values.
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
This commit modifies ompi's process list group object to support a
sentinel value for non-existant ompi_proc_t objects. The sentinel was
chosen to be the negative of the opal_process_name_t of the associated
ompi_proc_t. This takes advantage of the fact that on most (all?)
systems the top bit of a user-space pointer is never set. If this
changes then a new sentinel will be needed.
In addition this commit modifies the way ompi_mpi_comm_world is
initialized to fill in the group with sentinel values if the number of
processes exceeds the new add_procs behavior cutoff.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds an opal hash table to keep track of mapping between
process identifiers and ompi_proc_t's. This hash table is used by the
ompi_proc_by_name() function to lookup (in O(1) time) a given
process. This can be used by a BTL or other component to get a
ompi_proc_t when handling an incoming message from an as yet unknown
peer.
Additionally, this commit adds a new MCA variable to control the new
add_procs behavior: mpi_add_procs_cutoff. If the number of ranks in
the process falls below the threshold a ompi_proc_t is created for
every process. If the number of ranks is above the threshold then a
ompi_proc_t is only created for the local rank. The code needed to
generate additional ompi_proc_t's for a communicator is not yet
complete.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit contains the following changes:
- pml/ob1: use the bml accessor function when requesting a bml
endpoint. this will ensure that bml endpoints are only created when
needed. for example, a bml endpoint is not requested and not
allocated when receiving an eager message from a peer.
- pml/ob1: change the pml_procs array in the ob1 communicator to a
proc pointer array. at the cost of a single level of extra
redirection this will allow us to allocate pml procs on demand.
- pml/ob1: add an accessor function to access the pml proc structure
for a given peer. this function will allocate the proc if it
doesn't already exist.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit contains the following changes:
- bml: add a function to add a single process. this function is
intended to remove the need to maintain a opal_bitmap_t as it is
irrelevant for a single proc. BTLs will need to be updated to
either 1) ignore the return code from opal_bitmap_set_bit or not
call the function if the reachability bitmap is NULL.
- bml: add an inline accessor function for getting the bml endpoint
for a peer proc. this function will either 1) return the cached bml
endpoint, or 2) create the endpoint and call add_proc will all
available BTL modules.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
As of v15.7, the PGI Fortran compiler does not properly support how
Open MPI uses the "USE ... ONLY" Fortran syntax to include modules
with conflicting symbol definitions (interestingly, pgfortran only has
a problem with this when compiling with -g).
In short, OMPI uses "USE :: module_aaa, ONLY: foo" and "USE ::
module_bbb, ONLY: bar" to use modules aaa and bbb, even though they
contain conflicting definitions for some symbols. However, the use of
the ONLY clause should preclude the inclusion of the conflicting
symbols -- as the word implies, it should direct the compiler to
*only* use the symbols identified by the clause (i.e., foo and bar, in
this example).
This commit adds a configure test for this capability. If the
compiler fails to build a simple test that mimics this behavior, then
disable the mpi_f08 bindings.
Fixesopen-mpi/ompi#857
After long debugging, I found last week the reason this optimization originally broke
some hdf5 tests. We now pass the hdf5 test suite with the optimization being actively used.
Specifically:
- reduce the number of realloc's and malloc's by moving
some arrays out of the cycle loop, if we know that there
size is not changing
- store the rank of the aggregator in a separate variable to avoid
continuous dereferencing
- change the wait_all logic in write_all to use a fix number of requests
(even if they are MPI_REQUEST_NULL)
- fix the timing to considere the two initial allgather and the one
allgatherv operation to be a part of it
- add more comments.
- MPI_Compare_and_swap
- MPI_Fetch_and_op
- MPI_Raccumulate
- MPI_Win_detach
Thanks to Michael Knobloch and Takahiro Kawashima for bringing this
to our attention
Bring Slurm PMI-1 component online
Bring the s2 component online
Little cleanup - let the various PMIx modules set the process name during init, and then just raise it up to the ORTE level. Required as the different PMI environments all pass the jobid in different ways.
Bring the OMPI pubsub/pmi component online
Get comm_spawn working again
Ensure we always provide a cpuset, even if it is NULL
pmix/cray: adjust cray pmix component for pmix
Make changes so cray pmix can work within the integrated
ompi/pmix framework.
Bring singletons back online. Implement the comm_spawn operation using pmix - not tested yet
Cleanup comm_spawn - procs now starting, error in connect_accept
Complete integration
In an abort situation, just bail out immediately -- don't try to
invoke any atexit()/on_exit()-registered functions.
This is similar rationale to
open-mpi/ompi@17846411c3.
Ensure to define ompi/pompi versions for platforms that don't have
weak symbols. Also make fortran/mpif-h/profile build a separate
sizeof library, just like fortran/mpifh-h does.
Per http://www.open-mpi.org/community/lists/devel/2015/08/17775.php,
some compilers don't like it when there's a .f90 file that only
contains comments (and no actual Fortran code). So if OMPI determines
that the Fortran compiler does not support enough Fortran mojo to
support MPI_SIZEOF, generate at least one dummy Fortran subroutine
that can be compiled in an otherwise barren Fortran landscape that is
devoid of life and hope.
This pull request adds an arraylist of type Buffer to
the Request class. Whenever a request object is created
that has associated buffers, the buffers should be added
to this array list so the java garbage collector does
not dispose of the buffers prematurely.
This is a more robust expansion on the idea first proposed by
@ggouaillardet
Fixes#369
Signed-off-by: Nathaniel Graham <ngraham@lanl.gov>
Portals4 supports atomic ops on datatypes less than or equal to
max_fetch_atomic_size bytes. This commit fixes a bug that required
the datatype to be less than max_fetch_atomic_size bytes.
Fortran uses objects (ompi_f08_mpi_comm_world, mpi_fortran_bottom,, ...) that are defined in C.
Some compilers have different requirements on how these objects should be aligned.
Smaller alignment in C can lead to several confusing warnings from the linker, so try to
find the alignment expected by Fortran compiler, and inform the C compiler.
The libmpi_*.la fortran libraries make some direct calls to
libopen-pal.la. In many (most?) cases, having libmpi_* link
against libmpi is sufficient (because libmpi pulls in libopen-pal).
But when building RPMs on SLES, some compiler/linker flags are used
that seem to make this implicit linking not sufficient -- we get
missing opal symbols when creating libmpi_mpifh.la. So link in
open-pal directly (vs. indirectly), which solves the problem.
On some OSs (e.g., Ubuntu 14.04.2 LTS), the linker is configured such
that the symbols of library dependencies are not available to the
application. Hence, we need to explicitly list such dependencies when
creating the executable.
For this commit, these tests are use OPAL function calls, so we must
explicitly link in libopen-pal.so.
- make the internal structure follow the Open MPI naming convention
- provide a single flag/macro which controls the compilation/utilization of this
feature, to avoid that somebody using this has to modify every single
fcoll component. A configure option could be added later if desired.
