We have been getting several requests for new collectives that need to be inserted in various places of the MPI layer, all in support of either checkpoint/restart or various research efforts. Until now, this would require that the collective id's be generated at launch. which required modification
s to ORTE and other places. We chose not to make collectives reusable as the race conditions associated with resetting collective counters are daunti
ng.
This commit extends the collective system to allow self-generation of collective id's that the daemons need to support, thereby allowing developers to request any number of collectives for their work. There is one restriction: RTE collectives must occur at the process level - i.e., we don't curren
tly have a way of tagging the collective to a specific thread. From the comment in the code:
* In order to allow scalable
* generation of collective id's, they are formed as:
*
* top 32-bits are the jobid of the procs involved in
* the collective. For collectives across multiple jobs
* (e.g., in a connect_accept), the daemon jobid will
* be used as the id will be issued by mpirun. This
* won't cause problems because daemons don't use the
* collective_id
*
* bottom 32-bits are a rolling counter that recycles
* when the max is hit. The daemon will cleanup each
* collective upon completion, so this means a job can
* never have more than 2**32 collectives going on at
* a time. If someone needs more than that - they've got
* a problem.
*
* Note that this means (for now) that RTE-level collectives
* cannot be done by individual threads - they must be
* done at the overall process level. This is required as
* there is no guaranteed ordering for the collective id's,
* and all the participants must agree on the id of the
* collective they are executing. So if thread A on one
* process asks for a collective id before thread B does,
* but B asks before A on another process, the collectives will
* be mixed and not result in the expected behavior. We may
* find a way to relax this requirement in the future by
* adding a thread context id to the jobid field (maybe taking the
* lower 16-bits of that field).
This commit includes a test program (orte/test/mpi/coll_test.c) that cycles 100 times across barrier and modex collectives.
This commit was SVN r32203.
Change the priority of comm_failure and job_termination events to ensure we process final messages prior to terminating. Check for termination conditions when processing proc termination events as we may order proc termination when the daemon gets an exit command, but we can't see the proc actually terminate until we get out of that message event.
Jeff: probably easiest to review this by testing. I tested it under both Slurm and rsh on v1.7.5 as well as trunk
cmr=v1.7.5:reviewer=jsquyres:subject=resolve event priorities during VM shutdown
This commit was SVN r31042.
Fix comm_spawn on a single host - with the new default mapping scheme, we were incorrectly computing the number of procs to put on the node.
Refs trac:4003
This commit was SVN r30033.
The following Trac tickets were found above:
Ticket 4003 --> https://svn.open-mpi.org/trac/ompi/ticket/4003
Fix two problems that surfaced when using direct launch under SLURM:
1. locally store our own data because some BTLs want to retrieve
it during add_procs rather than use what they have internally
2. cleanup MPI_Abort so it correctly passes the error status all
the way down to the actual exit. When someone implemented the
"abort_peers" API, they left out the error status. So we lost
it at that point and *always* exited with a status of 1. This
forces a change to the API to include the status.
cmr:v1.7.3:reviewer=jsquyres:subject=Fix MPI_Abort and modex_recv for direct launch
This commit was SVN r29405.
The intercomm "merge" function can create a linkage between procs that was not reflected anywhere in a modex, and so at least some of the procs in the resulting communicator don't know how to talk to some of the new communicator's peers.
For example, consider the case where:
1. parent job A comm_spawns a process (job B) - these processes exchange modex and can communicate
2. parent job A now comm_spawns another process (job C) - again, these can communicate, but the proc in C knows nothing of B
3. do an intercomm merge across the communicators created by the two comm_spawns. This puts B and C into the same communicator, but they know nothing about how to talk to each other as they were not involved in any exchange of contact info. Hence, collectives on that communicator now fail.
This fix adds an API to the ompi/dpm framework that (a) exchanges the modex info across the procs in the merge to ensure all procs know how to communicate, and (b) calls add_procs to give the btl's a chance to select transports to any new procs.
cmr:v1.7.3:reviewer=jsquyres
This commit was SVN r29166.
The following Trac tickets were found above:
Ticket 2904 --> https://svn.open-mpi.org/trac/ompi/ticket/2904
* paccept - establish a persistent listening port for async connect requests
* pconnect - async connect to remote process that has posted a paccept port. Provides a timeout mechanism, and allows the underlying implementation to retry until timeout
* pclose - shuts down a prior paccept posting
Includes example programs paccept.c and pconnect.c in orte/test/mpi. New MPI extension interfaces coming...
This commit was SVN r29063.
*** THIS RFC INCLUDES A MINOR CHANGE TO THE MPI-RTE INTERFACE ***
Note: during the course of this work, it was necessary to completely separate the MPI and RTE progress engines. There were multiple places in the MPI layer where ORTE_WAIT_FOR_COMPLETION was being used. A new OMPI_WAIT_FOR_COMPLETION macro was created (defined in ompi/mca/rte/rte.h) that simply cycles across opal_progress until the provided flag becomes false. Places where the MPI layer blocked waiting for RTE to complete an event have been modified to use this macro.
***************************************************************************************
I am reissuing this RFC because of the time that has passed since its original release. Since its initial release and review, I have debugged it further to ensure it fully supports tests like loop_spawn. It therefore seems ready for merge back to the trunk. Given its prior review, I have set the timeout for one week.
The code is in https://bitbucket.org/rhc/ompi-oob2
WHAT: Rewrite of ORTE OOB
WHY: Support asynchronous progress and a host of other features
WHEN: Wed, August 21
SYNOPSIS:
The current OOB has served us well, but a number of limitations have been identified over the years. Specifically:
* it is only progressed when called via opal_progress, which can lead to hangs or recursive calls into libevent (which is not supported by that code)
* we've had issues when multiple NICs are available as the code doesn't "shift" messages between transports - thus, all nodes had to be available via the same TCP interface.
* the OOB "unloads" incoming opal_buffer_t objects during the transmission, thus preventing use of OBJ_RETAIN in the code when repeatedly sending the same message to multiple recipients
* there is no failover mechanism across NICs - if the selected NIC (or its attached switch) fails, we are forced to abort
* only one transport (i.e., component) can be "active"
The revised OOB resolves these problems:
* async progress is used for all application processes, with the progress thread blocking in the event library
* each available TCP NIC is supported by its own TCP module. The ability to asynchronously progress each module independently is provided, but not enabled by default (a runtime MCA parameter turns it "on")
* multi-address TCP NICs (e.g., a NIC with both an IPv4 and IPv6 address, or with virtual interfaces) are supported - reachability is determined by comparing the contact info for a peer against all addresses within the range covered by the address/mask pairs for the NIC.
* a message that arrives on one TCP NIC is automatically shifted to whatever NIC that is connected to the next "hop" if that peer cannot be reached by the incoming NIC. If no TCP module will reach the peer, then the OOB attempts to send the message via all other available components - if none can reach the peer, then an "error" is reported back to the RML, which then calls the errmgr for instructions.
* opal_buffer_t now conforms to standard object rules re OBJ_RETAIN as we no longer "unload" the incoming object
* NIC failure is reported to the TCP component, which then tries to resend the message across any other available TCP NIC. If that doesn't work, then the message is given back to the OOB base to try using other components. If all that fails, then the error is reported to the RML, which reports to the errmgr for instructions
* obviously from the above, multiple OOB components (e.g., TCP and UD) can be active in parallel
* the matching code has been moved to the RML (and out of the OOB/TCP component) so it is independent of transport
* routing is done by the individual OOB modules (as opposed to the RML). Thus, both routed and non-routed transports can simultaneously be active
* all blocking send/recv APIs have been removed. Everything operates asynchronously.
KNOWN LIMITATIONS:
* although provision is made for component failover as described above, the code for doing so has not been fully implemented yet. At the moment, if all connections for a given peer fail, the errmgr is notified of a "lost connection", which by default results in termination of the job if it was a lifeline
* the IPv6 code is present and compiles, but is not complete. Since the current IPv6 support in the OOB doesn't work anyway, I don't consider this a blocker
* routing is performed at the individual module level, yet the active routed component is selected on a global basis. We probably should update that to reflect that different transports may need/choose to route in different ways
* obviously, not every error path has been tested nor necessarily covered
* determining abnormal termination is more challenging than in the old code as we now potentially have multiple ways of connecting to a process. Ideally, we would declare "connection failed" when *all* transports can no longer reach the process, but that requires some additional (possibly complex) code. For now, the code replicates the old behavior only somewhat modified - i.e., if a module sees its connection fail, it checks to see if it is a lifeline. If so, it notifies the errmgr that the lifeline is lost - otherwise, it notifies the errmgr that a non-lifeline connection was lost.
* reachability is determined solely on the basis of a shared subnet address/mask - more sophisticated algorithms (e.g., the one used in the tcp btl) are required to handle routing via gateways
* the RML needs to assign sequence numbers to each message on a per-peer basis. The receiving RML will then deliver messages in order, thus preventing out-of-order messaging in the case where messages travel across different transports or a message needs to be redirected/resent due to failure of a NIC
This commit was SVN r29058.
additional functionality. Rationale (refs trac:3422):
* Normal MPI applications only ever use the MPI API. Hence, -lmpi is
sufficient (they'll never directly call ORTE or OPAL
functions). This is arguably the most common case.
* That being said, we do have some test programs (e.g., those in
orte/test/mpi) that call MPI functions but also call ORTE/OPAL
functions. I've also written the occasional MPI test program that
calls opal_output, for example (there even might be a few tests in
the IBM test suite that directly call ORTE/OPAL functions).
* Even though this is not a common case, these applications should
also compile/link with mpicc.
* So we should add a --openmpi:linkall option that will also link
in whatever is necessary to call ORTE/OPAL functions
* Yes, we could hard-code "-lopen-rte -lopen-pal" in Makefiles, but
we do reserve the right to change those library names and/or add
others someday, so it's better to abstract out the names and let
the wrapper supply whatever is necessary.
* ORTE programs, however, are different. They almost always call OPAL
functions (e.g., if they want to send a message, they must use the
OPAL DSS). As such, it seems like the ORTE programs should always
link in OPAL.
Therefore:
* Add undocumented --openmpi:linkall flag to the wrapper compilers.
See the comment in opal_wrapper.c for an explanation of what it
does. This flag is only intended for Open MPI developers -- not
end users. That's why it's undocumented.
* Update orte/test/mpi/Makefile.am to add --openmpi:linkall
* Make ortecc/ortec++'s wrapper data text files always explicitly
link in libopen-pal
This commit was SVN r27670.
The following SVN revision numbers were found above:
r27668 --> open-mpi/ompi@cf845897aa
The following Trac tickets were found above:
Ticket 3422 --> https://svn.open-mpi.org/trac/ompi/ticket/3422
following:
* Provides a fixed number of resource slots (i.e., "hotel rooms").
* Allows one thing to occupy a resource slot at a time (i.e., each
hotel room can have an occupant check in to that room).
* Resource slots can be vacated at any time (i.e., occupants can
voluntarily check out of their hotel room).
* Resource slots can be occupied for a specific maximum amount of
time. If that time expires, the occupant is forcibly evicted and
the upper layer is notified via (libevent) callback (i.e., the maid
will kick an occupant of out of their room when their reservation
is over).
This class can be to be used for things like retransmission schemes
for unreliable transports. For example, a message sent on an
unreliable transport can be checked in to a hotel room. If an ACK for
that message is received, the message can be checked out. But if the
ACK is never received, the message will eventually be evicted from its
room and the upper layer will be notified that the message failed to
check out in time (i.e., that an ACK for that message was not received
in time).
Code using this class is currently being developed off-trunk, but will
be coming to SVN soon.
This commit was SVN r27067.
"num_app_ctx" - the number of app_contexts in the job
"first_rank" - the MPI rank of the first process in each app_context
"np" - the number of procs in each app_context
Still need clarification on the MPI_Init portion of the ticket. Specifically, does the ticket call for returning an error is someone calls MPI_Init more than once in a program? We set a flag to tell us that we have been initialized, but currently never check it.
This commit was SVN r27005.
