On x86_64 reading a 128-bit value requires multiple instructions.
Under some conditions if the counted pointer counter is read before
the item pointer the fifo can be left in an inconsistent state. This
commit forces the read of the counter to always be read first.
The fifo does not appear to suffer from the same race.
It is possible the compiler can reorder the read of the head item and
the head itself. This could lead to a situation where the item
returned was not really the head item.
This commit adds a new class: opal_fifo.h. The new class has atomic, non-atomic,
and opal_using_threads() conditoned routines. It should be used when first-in
first-out is required and should perform much better than using locks and an
opal_list_t. Like with opal_lifo_t there are two versions of the atomic
implementation: 128-bit compare-and-swap, and spin-locked. More implementations
can be added later (LL/SC comes to mind).
This commit also adds a unit test for the opal_fifo_t class. This test verifies
the fifo implementation when using multiple threads.
- Rename opal_atomic_lifo_t to opal_lifo_t to reflect both atomic and
non-atomic usage. Added new routines (opal_lifo_*_st) for non-atomic
usage as well as routines conditioned off opal_using_threads(). The
atomic versions are always thread safe and the non-atomic are always
not thread safe.
- Add a new atomic lifo implementation that makes use of 128-bit
compare-and-swap. The new implementation should scale better with
larger numbers of threads.
- Add threading unit test for opal_lifo_t.
We recognize that this means other users of OPAL will need to "wrap" the opal_process_name_t if they desire to abstract it in some fashion. This is regrettable, and we are looking at possible alternatives that might mitigate that requirement. Meantime, however, we have to put the needs of the OMPI community first, and are taking this step to restore hetero and SPARC support.
These two macros set the prefix for the OPAL and ORTE libraries,
respectively. Specifically, the OPAL library will be named
libPREFIXopen-pal.la and the ORTE library will be named
libPREFIXopen-rte.la.
These macros must be called, even if the prefix argument is empty.
The intent is that Open MPI will call these macros with an empty
prefix, but other projects (such as ORCM) will call these macros with
a non-empty prefix. For example, ORCM libraries can be named
liborcm-open-pal.la and liborcm-open-rte.la.
This scheme is necessary to allow running Open MPI applications under
systems that use their own versions of ORTE and OPAL. For example,
when running MPI applications under ORTE, if the ORTE and OPAL
libraries between OMPI and ORCM are not identical (which, because they
are released at different times, are likely to be different), we need
to ensure that the OMPI applications link against their ORTE and OPAL
libraries, but the ORCM executables link against their ORTE and OPAL
libraries.
the OPAL and ORTE libraries. This is required by projects such as ORCM
that have their own ORTE and OPAL libraries in order to avoid library
confusion. By renaming their version of the libraries, the OMPI
applications can correctly dynamically load the correct one for their
build."
This reverts commit 63f619f871.
Revert "OPAL: drop dead with core on bad flow. rarely happens with helloworld on large scale."
This reverts commit 86f1d5af3e.
Will be reconsidered via RFC as it represents a significant change in behavior
1. It's actually hashing now, whereas the old OPAL hash table was not. Thus, it is a bug fix for and, as such, should be included in the 1.8 series.
2. It is dynamic and can grow and shrink the number of buckets in accordance with job size, whereas the old OPAL hash table had a fixed number of buckets which resulted in poor retrieval performance at large scale.
This scheme has been deployed in the field on very large H.P./Mellanox systems and has been demonstrated to significantly decrease job start-up time (~ 20% improvement) when launching applications directly with srun in SLURM environments. However, neither SLURM nor direct launch are prerequisites to take advantage of this change as any entity that utilizes OPAL hash table objects can benefit (at least partially) from this contribution.
With --enable-memchecker builds, use calloc(3) for OBJ_NEW instead of
malloc(3). This cuts down on a lot of valgrind/memory checker false
positive output.
Also make a minor change in the valgrind configure.m4; have it assign
0xf to a char. The prior assignment (of 0xff) was warning about an
overflow. This didn't really matter, but we might as well make the
test not have a gratuitious warning in it.
Decrease the hash table size when an element is removed
cmr=v1.8.2:ticket=trac:4872
This commit was SVN r32566.
The following Trac tickets were found above:
Ticket 4872 --> https://svn.open-mpi.org/trac/ompi/ticket/4872
use HAVE_STRINGS_H to protect <strings.h> include
cmr=v1.8.2:reviewer=rhc:ticket=trac:4853
This commit was SVN r32510.
The following Trac tickets were found above:
Ticket 4853 --> https://svn.open-mpi.org/trac/ompi/ticket/4853
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.
turned out to be a nightmare, as the pointers to the classes are
located in shared libraries memory areas, and are not accesible
after the shared library is unloaded. Thus, OPAL cannot cleanup
the left-over classes from the other shared libraries.
This commit was SVN r32248.
The following SVN revision numbers were found above:
r32236 --> open-mpi/ompi@59017433e1
Need to remove the items of the list to avoid an assert in debug builds.
cmr=v1.8.2:ticket=trac:4628
This commit was SVN r31769.
The following SVN revision numbers were found above:
r31764 --> open-mpi/ompi@13fd6ae774
The following Trac tickets were found above:
Ticket 4628 --> https://svn.open-mpi.org/trac/ompi/ticket/4628
list
This commit updates the behavior of opal_free_list_t to match the
behavior of ompi_free_list_t. opal_free_list_t constructed items
placed on the free list but never destructed them.
cmr=v1.8.2:reviewer=jsquyres
This commit was SVN r31764.
r30086: make sure that a super item is constructed properly).
Refs trac:4035
This commit was SVN r30090.
The following SVN revision numbers were found above:
r30086 --> open-mpi/ompi@d1c63f878e
The following Trac tickets were found above:
Ticket 4035 --> https://svn.open-mpi.org/trac/ompi/ticket/4035
Thanks to Tetsuya Mishima for identifying the problem and providing the patch!
cmr=v1.7.4:reviewer=jsquyres:subject=Fix LAMA mapper for PGI compilers
This commit was SVN r30086.
more:
- Remove OPAL_ENABLE_MULTI_THREADS, since it didn't really do anything
correctly. Opal always has threads enabled at this point.
- Remove OMPI_ENABLE_PROGRESS_THREADS, since this hasn't worked in
8 years and it has performance issues we'll never be able to
overcome. Note that we have plans for re-adding async progress, using
a hybrid protocol of async and sync sends.
- OMPI_ENABLE_THREAD_MULTIPLE now determines whether the thread lock
macros do the check or not.
- Condition variables are ALWAYS polling right now, which fixes the thread
live-lock currently found when THREAD_MULTIPLE is turned on.
This commit was SVN r29891.
(since this is used in the fast path) for when you ''know'' that there
will be a room available:
* Don't do the last_unoccupied_room check
* Return void
This commit was SVN r28757.
Chose the crc32 function present in opal/util/crc.c as the hash function. The
performance should be sufficient for most cases. If not we can always change
the function again.
This commit was SVN r28629.
actually care if opal_pointer_array is limited to handle_max already passes
that in as the max_size during init, so don't need it there. The arch
constant was a bit more difficult, so pass that in during MPI init and
leave empty otherwise.
This is to help with the effort to allow building ompi against an external
opal or orte.
This commit was SVN r27817.
