1
1
Граф коммитов

62 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
5e6536eeda Ensure that mpirun properly accounts for itself when exiting without reply.
Move some debug output around so it is always seen.

This commit was SVN r20026.
2008-11-20 18:55:59 +00:00
Ralph Castain
9ba78f6e5f Ensure exit-no-reply gets relayed to downstream orteds prior to exiting ourselves
This commit was SVN r20021.
2008-11-18 14:54:52 +00:00
Ralph Castain
f54fda489e This is a first step towards supporting fully-routed OOB communications:
1. remove direct routed module (hooray!)

2. add radix tree routed module (binomial remains default)

3. remove duplicate data storage - orteds were storing nidmap and pidmap data in odls, everyone else in ess

4. add ess APIs to update nidmap, add new pidmap - used only by orteds for MPI-2 support

5. modify code to eliminate multiple calls to orte_routed.update_route that recreated info already in ess pidmap. Add ess API to lookup that info instead. Modify routed modules to utilize that capability

6. setup new ability to shutdown orteds without sending back an "ack" message to mpirun - not utilized yet, will require some changes to plm terminate_orteds functions in managed environments (coming soon)

Initial tests indicating that fully routing comm via defined routing trees may not actually have a significant cost for operations like IB QP setup. More tests required to confirm.

This will require an autogen...

This commit was SVN r19866.
2008-10-31 21:10:00 +00:00
Ralph Castain
48c3de1865 Fix a problem in the plm "failed to start" code observed by Jeff. When we are unable to launch to a specific node because it doesn't exist or is down, the system would hang and/or segv. The reason for the hang was that we were "firing" the orted exit trigger prior to its timer event being defined - thus "locking" that one-shot and preventing it from firing when we actually were ready to use it.
The segv was caused by the fact that we don't really know which daemon failed to start (at least, in most cases), so we didn't set a pointer to the aborted proc object. All we really wanted, though, was to ensure that mpirun returned a non-zero exit status, so the fix was to simply return the default error status.

This commit was SVN r19754.
2008-10-16 14:21:37 +00:00
Ralph Castain
9b8473fdbf Cleanup orted cmd line - we don't need to pass nodenames, and shouldn't pass heartbeat unless the orted is going to use it. This helps shorten the cmd line for future use.
Cleanup when an orted actually opens the PLM. Unfortunately, some unmentionable people are pushing head node environs out to remote nodes, causing the daemons to think they are the HNP. This helps prevent the confusion.

This commit was SVN r19518.
2008-09-08 15:45:11 +00:00
Ralph Castain
63c33a9c32 Some minor updates to the locking system changes. Remove obsolete locks. Ensure the trigger event objects do not get deconstructed until the very end to avoid possible problems due to race conditions. Route all orted abnormal term tests through the trigger.
This commit was SVN r19172.
2008-08-06 11:31:06 +00:00
Ralph Castain
be02211b4f Modify the wakeup system to make it more Windows-friendly. This allows Shiqing to consolidate the Windows-specific modifications into one location, and generalizes the wakeup procedure in case we hit other system-specific requirements.
This needs some soak time to ensure we haven't opened any race conditions. I tried to loop everything in the shutdown procedure through that trigger event call to ensure it all goes through the one-time locks as it did before so that someone hitting ctrl-c when we are already shutting down shouldn't cause problems. Just want to let people use it for awhile to verify.

This commit was SVN r19159.
2008-08-05 15:09:29 +00:00
Ralph Castain
f1114b4144 Upgrade the ability of orterun to deal with cmd line MCA params that are passed to the orteds. Help reduce the size of the cmd line by eliminating duplicates where possible, and alert to duplicate entries that can cause problems.
Add comments to both orterun and orted code explaining why we take a snapshot of the local environment and apply it to the local procs when they are spawned.

This commit was SVN r18842.
2008-07-08 22:36:39 +00:00
George Bosilca
07cb54995b Reactivate the daemon spin from the command line.
This commit was SVN r18794.
2008-07-02 01:46:58 +00:00
Ralph Castain
3b5e80fa61 Shift responsibility for preconnecting the oob to the orte routed framework, which is the only place that knows what needs to be done. Only the direct module will actually do anything - it uses the same algo as the original preconnect function.
This commit was SVN r18677.
2008-06-19 13:48:26 +00:00
Ralph Castain
0532d799d6 Complete implementation of the --without-rte-support configure option. Working with Brian, this has been tested on RedStorm.
Some minor changes to help facilitate debugger support so that both mpirun and yod can operate with it. Still to be completed.

This commit was SVN r18664.
2008-06-18 03:15:56 +00:00
Ralph Castain
9613b3176c Effectively revert the orte_output system and return to direct use of opal_output at all levels. Retain the orte_show_help subsystem to allow aggregation of show_help messages at the HNP.
After much work by Jeff and myself, and quite a lot of discussion, it has become clear that we simply cannot resolve the infinite loops caused by RML-involved subsystems calling orte_output. The original rationale for the change to orte_output has also been reduced by shifting the output of XML-formatted vs human readable messages to an alternative approach.

I have globally replaced the orte_output/ORTE_OUTPUT calls in the code base, as well as the corresponding .h file name. I have test compiled and run this on the various environments within my reach, so hopefully this will prove minimally disruptive.

This commit was SVN r18619.
2008-06-09 14:53:58 +00:00
Ralph Castain
7bee71aa59 Fix a potential, albeit perhaps esoteric, race condition that can occur for fast HNP's, slow orteds, and fast apps. Under those conditions, it is possible for the orted to be caught in its original send of contact info back to the HNP, and thus for the progress stack never to recover back to a high level. In those circumstances, the orted can "hang" when trying to exit.
Add a new function to opal_progress that tells us our recursion depth to support that solution.

Yes, I know this sounds picky, but good ol' Jeff managed to make it happen by driving his cluster near to death...

Also ensure that we declare "failed" for the daemon job when daemons fail instead of the application job. This is important so that orte knows that it cannot use xcast to tell daemons to "exit", nor should it expect all daemons to respond. Otherwise, it is possible to hang.

After lots of testing, decide to default (again) to slurm detecting failed orteds. This proved necessary to avoid rather annoying hangs that were difficult to recover from. There are conditions where slurm will fail to launch all daemons (slurm folks are working on it), and yet again, good ol' Jeff managed to find both of them.

Thanks you Jeff! :-/

This commit was SVN r18611.
2008-06-06 19:36:27 +00:00
Ralph Castain
95578b0528 Fix single-node operations so that the HNP correctly exits when the job completes
This commit was SVN r18556.
2008-06-03 14:23:04 +00:00
Ralph Castain
b456fb2d42 Upgrade the node/orted failure detection code to cover all environments. Use the native environment's capabilities where possible - e.g., SLURM detects orted failure and can report it. Elsewhere, use a heartbeat system to detect orted failure - e.g., for TM and rsh. Heart rate is set via mca param. The HNP checks for callback every 2*heartrate, declares orted failure if not seen in last 2*heartrate time.
Also detect orted failed-to-start by setting timeout on launch. Currently only used in TM launcher.

Neither detection is enabled by default, but are only active if heartrate is set and/or launch timeout is set. Exception for SLURM as orted failure is always detected and reported.

More info to come on devel list.

This commit was SVN r18555.
2008-06-02 21:46:34 +00:00
Ralph Castain
72530f8fed Cleanly handle the failed start of an orted, or its unexpected failure after start. This commit will allow mpirun to exit cleanly when this occurs, and does a best-effort attempt to cleanup the mess. However, it still has two unresolved issues that need to be eventually addressed:
1. it depends upon the ability of the native environment to alert us that the orted has died/failed to start. I have included that support for SLURM, but other environments need to be done.

2. for some yet-to-be-determined reason, the message that tells the remaining daemons to "die" isn't getting out of the RML, even though no obvious blockage is standing in the way. Work will continue on resolving that problem. For now, the orteds appear to be exiting on their own quite nicely when they see their HNP "lifeline" disappear.

This represents the best-available fix for ticket #221 so I am closing that ticket at this time.

This commit was SVN r18536.
2008-05-29 13:38:27 +00:00
Jeff Squyres
e7ecd56bd2 This commit represents a bunch of work on a Mercurial side branch. As
such, the commit message back to the master SVN repository is fairly
long.

= ORTE Job-Level Output Messages =

Add two new interfaces that should be used for all new code throughout
the ORTE and OMPI layers (we already make the search-and-replace on
the existing ORTE / OMPI layers):

 * orte_output(): (and corresponding friends ORTE_OUTPUT,
   orte_output_verbose, etc.)  This function sends the output directly
   to the HNP for processing as part of a job-specific output
   channel.  It supports all the same outputs as opal_output()
   (syslog, file, stdout, stderr), but for stdout/stderr, the output
   is sent to the HNP for processing and output.  More on this below.
 * orte_show_help(): This function is a drop-in-replacement for
   opal_show_help(), with two differences in functionality:
   1. the rendered text help message output is sent to the HNP for
      display (rather than outputting directly into the process' stderr
      stream)
   1. the HNP detects duplicate help messages and does not display them
      (so that you don't see the same error message N times, once from
      each of your N MPI processes); instead, it counts "new" instances
      of the help message and displays a message every ~5 seconds when
      there are new ones ("I got X new copies of the help message...")

opal_show_help and opal_output still exist, but they only output in
the current process.  The intent for the new orte_* functions is that
they can apply job-level intelligence to the output.  As such, we
recommend that all new ORTE and OMPI code use the new orte_*
functions, not thei opal_* functions.

=== New code ===

For ORTE and OMPI programmers, here's what you need to do differently
in new code:

 * Do not include opal/util/show_help.h or opal/util/output.h.
   Instead, include orte/util/output.h (this one header file has
   declarations for both the orte_output() series of functions and
   orte_show_help()).
 * Effectively s/opal_output/orte_output/gi throughout your code.
   Note that orte_output_open() takes a slightly different argument
   list (as a way to pass data to the filtering stream -- see below),
   so you if explicitly call opal_output_open(), you'll need to
   slightly adapt to the new signature of orte_output_open().
 * Literally s/opal_show_help/orte_show_help/.  The function signature
   is identical.

=== Notes ===

 * orte_output'ing to stream 0 will do similar to what
   opal_output'ing did, so leaving a hard-coded "0" as the first
   argument is safe.
 * For systems that do not use ORTE's RML or the HNP, the effect of
   orte_output_* and orte_show_help will be identical to their opal
   counterparts (the additional information passed to
   orte_output_open() will be lost!).  Indeed, the orte_* functions
   simply become trivial wrappers to their opal_* counterparts.  Note
   that we have not tested this; the code is simple but it is quite
   possible that we mucked something up.

= Filter Framework =

Messages sent view the new orte_* functions described above and
messages output via the IOF on the HNP will now optionally be passed
through a new "filter" framework before being output to
stdout/stderr.  The "filter" OPAL MCA framework is intended to allow
preprocessing to messages before they are sent to their final
destinations.  The first component that was written in the filter
framework was to create an XML stream, segregating all the messages
into different XML tags, etc.  This will allow 3rd party tools to read
the stdout/stderr from the HNP and be able to know exactly what each
text message is (e.g., a help message, another OMPI infrastructure
message, stdout from the user process, stderr from the user process,
etc.).

Filtering is not active by default.  Filter components must be
specifically requested, such as:

{{{
$ mpirun --mca filter xml ...
}}}

There can only be one filter component active.

= New MCA Parameters =

The new functionality described above introduces two new MCA
parameters:

 * '''orte_base_help_aggregate''': Defaults to 1 (true), meaning that
   help messages will be aggregated, as described above.  If set to 0,
   all help messages will be displayed, even if they are duplicates
   (i.e., the original behavior).
 * '''orte_base_show_output_recursions''': An MCA parameter to help
   debug one of the known issues, described below.  It is likely that
   this MCA parameter will disappear before v1.3 final.

= Known Issues =

 * The XML filter component is not complete.  The current output from
   this component is preliminary and not real XML.  A bit more work
   needs to be done to configure.m4 search for an appropriate XML
   library/link it in/use it at run time.
 * There are possible recursion loops in the orte_output() and
   orte_show_help() functions -- e.g., if RML send calls orte_output()
   or orte_show_help().  We have some ideas how to fix these, but
   figured that it was ok to commit before feature freeze with known
   issues.  The code currently contains sub-optimal workarounds so
   that this will not be a problem, but it would be good to actually
   solve the problem rather than have hackish workarounds before v1.3 final.

This commit was SVN r18434.
2008-05-13 20:00:55 +00:00
Ralph Castain
40904dd152 Add a binomial routed module - for now, still completely wires up the daemons, but that will be changed later.
Modify grpcomm xcast so it now uses the selected routed module - eliminates cross-wiring of xcast and routing paths. Suboptimal at the moment, but better implementation is on its way.

Cleanup ignore properties on the new routed components.

This commit was SVN r18377.
2008-05-05 22:32:25 +00:00
Ralph Castain
519c15f8af Fix direct and linear xcast modes
This commit was SVN r18359.
2008-05-02 14:30:07 +00:00
Ralph Castain
b2c73f6e11 Fix tree-spawn to work within the new modex system
This commit was SVN r18349.
2008-05-01 19:19:34 +00:00
Ralph Castain
3e55fe6f6d Fold in the revised modex scheme. Move the ompi_proc_t modex portions to the RTE level since the daemons already have that info. Provide each process with the equivalent of a "nidmap" - both a map of what nodes are in the job, and a map of which node each process is on. This enables the use of static ports, though that hasn't been turned "on" in this commit.
Update the rsh tree spawn capability so we spawn the next wave of daemons before launching our own local procs.

Add an ability to encode nodenames for large clusters with contiguous node name numbering schemes - this allows communication of all node names in a few bytes instead of tens-of-bytes/node.

This commit was SVN r18338.
2008-04-30 19:49:53 +00:00
Ralph Castain
4c2c6c9bd8 Ensure the pack/unpacks match for tree-spawn
This commit was SVN r18282.
2008-04-24 18:53:08 +00:00
Ralph Castain
09b6758f8c Pass the prefix dir to the remote orted when doing tree-based spawns
This commit was SVN r18280.
2008-04-24 18:38:24 +00:00
Josh Hursey
cc83d41ad9 Merge in tmp/jjh-scratch
{{{
 svn merge -r 18218:18240 https://svn.open-mpi.org/svn/ompi/tmp/jjh-scratch .
}}}

Contains:
 * Primarily a fix for a user reported problem where a cached file descriptor is causing a SIGPIPE on restart.
 * Cleanup some small memory leaks from using mca_base_param_env_var() - Thanks Jeff
 * Cleanup ORTE FT tool compilation in non-FT builds - Thanks Tim P.
 * Cleanup mpi interface with missplaced {{{OPAL_CR_ENTER_LIBRARY}}} - Thanks Terry
 * Some other sundry cleanup items all dealing with C/R functionality in the trunk.

This commit was SVN r18241.
2008-04-23 00:17:12 +00:00
Ralph Castain
eb27e4f23d Move the reissuing of the daemon recv to occur after the message actually gets processed. This ensures that we don't get multiple messages trying to be processed at the same time.
Add one more debug output to see where messages are heading

This commit was SVN r18183.
2008-04-16 20:41:00 +00:00
Ralph Castain
7c7304466c Add a binomial tree-based launch to ssh, turned "on" only when the plm_rsh_tree_spawned mca param is set to a non-zero value. This probably isn't a very optimized capability, but it does execute a tree-based launch that may scale better than linear at high node counts.
Add the daemon map capability to the ODLS to create and save a map of daemon vpid vs nodename from the launch message.

Cleanup a few places in the base plm launch support where we didn't adequately protect rml recv's from potentially executing sends.

This commit was SVN r18143.
2008-04-14 18:26:08 +00:00
Ralph Castain
3a0d09300b Fully implement the inbound binomial allgather for daemon-based collectives. Supports both modex and barrier operations.
Comm_spawn still uses the rank=0 method - shifting that algo to the daemons is under study.

This commit was SVN r18115.
2008-04-09 22:10:53 +00:00
Ralph Castain
6166278e18 Improve the scalability of the modex operation and fix a bug reported by Tim P
The bug was a race condition in the barrier operation that caused the barrier in MPI_Finalize to fail on very short programs.

Scalaiblity was improved by using the daemons to aggregate modex and barrier messages before sending them to the rank=0 proc. Improvement is proportional to ppn, of course, but there really wasn't a scaling problem at low ppn anyway. This modification also paves the way for better allgather operations since now all the data for each node is sitting at the daemon level, and the daemons are now aware that a collective operation on the OOB is underway (so they -can- participate in a collective of their own to support it).

Also added better diagnostics to map out the timing associated with MPI_Init - turned on by -mca orte_timing 1.

This commit was SVN r17988.
2008-03-27 15:17:53 +00:00
Jeff Squyres
763218e754 Fix #1253: default libevent to use select/poll and only use the other
mechanisms (such as epoll) if someone (ompi_mpi_init()) requests
otherwise.  See big comment in opal/event/event.c for a full
explanation.

This commit was SVN r17956.
2008-03-25 17:18:17 +00:00
Ralph Castain
4efddc7b0a Fix the allgather and allgather_list functions to avoid deadlocks at large node/proc counts. Violated the RML rules here - we received the allgather buffer and then did an xcast, which causes a send to go out, and is then subsequently received by the sender. This fix breaks that pattern by forcing the recv to complete outside of the function itself - thus, the allgather and allgather_list always complete their recvs before returning or sending.
Reogranize the grpcomm code a little to provide support for soon-to-come new grpcomm components. The revised organization puts what will be common code elements in the base to avoid duplication, while allowing components that don't need those functions to ignore them.

This commit was SVN r17941.
2008-03-24 20:50:31 +00:00
Ralph Castain
dc7f45dafd Remove the obsolete and largely unused orte_system_info structure. The only fields that were used in that struct were nodeid and nodename - these have been transferred to the orte_process_info structure.
Only one place used the user name field - session_dir, when formulating the name of the top-level directory. Accordingly, the code for getting the user's id has been moved to the session_dir code.

This commit was SVN r17926.
2008-03-23 23:10:15 +00:00
Jeff Squyres
314ab2c6e7 Update internal libevent to upstream (v1.4.2-rc + OMPI changes).
Greatly reduce the number of "foo" -> "opal_foo" symbol renames in the
libevent source, and instead greatly expand the event_rename.h file
that uses preprocessor macros to make all public symbols be
"opal_foo".

This commit was SVN r17923.
2008-03-23 12:33:04 +00:00
Ralph Castain
f8a10dfb93 Complete the fix of the orted vs mpirun race condition for finalizing. The darned mpirun is just too fast! Rather than try to slow it down, we set the orte_finalizing flag -prior- to telling mpirun the orted is leaving. This ensures we don't mistakenly declare the lifeline lost when mpirun leaves in a hurry.
This commit was SVN r17897.
2008-03-20 16:55:24 +00:00
Ralph Castain
27a73ad9ee Fix a race condition between the orteds and HNP that can cause the orteds to output the "lost lifeline" message.
This has been a long-time problem. I tried to reduce the problem by having the orteds tell the HNP they were finalizing, and having the HNP wait until all orteds had reported or we timed out.

What was observed was that all the orteds were correctly reporting that they are leaving, but the HNP is able to exit before the orteds, thus closing the orteds lifeline socket and generating the error output. This is caused by the fact that the orteds have to whack all remaining session directories, which includes that blasted monster shared memory file! Cleaning up the SM file can take quite a while.

The HNP doesn't have that problem as there is no SM file there! So it gets out first.

What we had done in the past to resolve that problem was put a little test in the OOB that checks to see if we are finalizing. If we are, then we ignore the lifeline connection being lost. That check was still in the code - however, we had lost the line in orte_finalize that set the flag!!

This commit was SVN r17893.
2008-03-20 13:30:51 +00:00
Ralph Castain
2ed0e60321 Bring some sanity to the exit code returned by mpirun. Ensure that we provide a non-zero code if something goes wrong, including someone exiting after calling mpi_init without calling mpi_finalize.
Jeff is preparing an (undoubtedly lengthy) explanation/matrix of how these codes are determined for the OMPI FAQ.

This commit was SVN r17879.
2008-03-19 19:00:51 +00:00
Ralph Castain
629b95a2fe Afraid this has a couple of things mixed into the commit. Couldn't be helped - had missed one commit prior to running out the door on vacation.
Fix race conditions in abnormal terminations. We had done a first-cut at this in a prior commit. However, the window remained partially open due to the fact that the HNP has multiple paths leading to orte_finalize. Most of our frameworks don't care if they are finalized more than once, but one of them does, which meant we segfaulted if orte_finalize got called more than once. Besides, we really shouldn't be doing that anyway.

So we now introduce a set of atomic locks that prevent us from multiply calling abort, attempting to call orte_finalize, etc. My initial tests indicate this is working cleanly, but since it is a race condition issue, more testing will have to be done before we know for sure that this problem has been licked.

Also, some updates relevant to the tool comm library snuck in here. Since those also touched the orted code (as did the prior changes), I didn't want to attempt to separate them out - besides, they are coming in soon anyway. More on them later as that functionality approaches completion.

This commit was SVN r17843.
2008-03-17 17:58:59 +00:00
Ralph Castain
841d0e5208 Cleanup an attribute warning - not sure which one to set or where it should go, so I'll leave that to someone more familiar with "attributes".
Ensure some debugging is only enabled when have_debug is set.

This commit was SVN r17681.
2008-03-03 16:06:47 +00:00
Ralph Castain
6450962d59 Add some debugging to the message event object.
Cleanup some no-longer-used values

This commit was SVN r17671.
2008-02-29 20:10:31 +00:00
Ralph Castain
a1eef0dd50 Fix a race condition in the orted recv/process procedure.
Thx to Tim P for spotting it

This commit was SVN r17666.
2008-02-29 15:18:45 +00:00
Ralph Castain
a585923de1 Silence some minor compiler warnings
This commit was SVN r17662.
2008-02-29 02:39:39 +00:00
Tim Prins
84b2099fe8 Remove the now-unused orte_value_array. As this is the last 'class' split between orte and ompi, remove the big comment about the split in ompi_bitmap.
Also, update some properties (source files should not be executeable...), and remove a couple unneeded inclusions of orte_proc_table.h

This commit was SVN r17655.
2008-02-28 21:39:42 +00:00
Ralph Castain
5e6928d710 Cleanup recursions in ORTE caused by processing recv'd messages that can cause the system to take action resulting in receipt of another message.
Basically, the method employed here is to have a recv create a zero-time timer event that causes the event library to execute a function that processes the message once the recv returns. Thus, any action taken as a result of processing the message occur outside of a recv.

Created two new macros to assist:

ORTE_MESSAGE_EVENT: creates the zero-time event, passing info in a new orte_message_event_t object

ORTE_PROGRESSED_WAIT: while waiting for specified conditions, just calls progress so messages can be recv'd.

Also fixed the failed_launch function as we no longer block in the orted callback function. Updated the error messages to reflect revision. No change in API to this function, but PLM "owners" may want to check their internal error messages to avoid duplication and excessive output.

This has been tested on Mac, TM, and SLURM.

This commit was SVN r17647.
2008-02-28 19:58:32 +00:00
George Bosilca
9d421bea2a Replace all occurences of orte_pointer_array by opal_pointer_array. Remove the
implementation of orte_pointer_array.

This commit was SVN r17636.
2008-02-28 05:32:23 +00:00
Ralph Castain
d70e2e8c2b Merge the ORTE devel branch into the main trunk. Details of what this means will be circulated separately.
Remains to be tested to ensure everything came over cleanly, so please continue to withhold commits a little longer

This commit was SVN r17632.
2008-02-28 01:57:57 +00:00
Josh Hursey
0bf61a1b84 Move in some accumulated small features and minor bug fixes for C/R support.
{{{
svn merge -r 16447:16475 https://svn.open-mpi.org/svn/ompi/tmp/jjh-fgs .
}}}

This commit was SVN r16478.
2007-10-17 13:47:36 +00:00
Jeff Squyres
423f23eb6a Fixes trac:1160. There is still some other problem in the OOB, but we
wanted to commit this to get wider testing.

This commit was SVN r16445.

The following Trac tickets were found above:
  Ticket 1160 --> https://svn.open-mpi.org/trac/ompi/ticket/1160
2007-10-15 15:41:36 +00:00
Jeff Squyres
fff1057597 We do ''not'' want the orted to yield the processor more than
necessary (it already uses blocking semantics while waiting for events
on fd's, so it's not taking more cycles from MPI applications than
necessary), or this can/will cause lengthy delays in orted processing.
The problem we most recently saw was with routed OOB messages getting
significantly delayed before being delivered to the target MPI
processes.  This was problematic for BTLs that use OOB messaging for
wireup.  There is a lengthy comment in orted_main.c that describes
this in more detail.

Setting the orted to not voluntarily call yield() prevents this bad
behavior.  The orted only runs in small bursts anyway (and blocks the
rest of the time), so this is not harmful to application performance.

This commit was SVN r16365.
2007-10-06 09:24:51 +00:00
Ralph Castain
54b2cf747e These changes were mostly captured in a prior RFC (except for #2 below) and are aimed specifically at improving startup performance and setting up the remaining modifications described in that RFC.
The commit has been tested for C/R and Cray operations, and on Odin (SLURM, rsh) and RoadRunner (TM). I tried to update all environments, but obviously could not test them. I know that Windows needs some work, and have highlighted what is know to be needed in the odls process component.

This represents a lot of work by Brian, Tim P, Josh, and myself, with much advice from Jeff and others. For posterity, I have appended a copy of the email describing the work that was done:

As we have repeatedly noted, the modex operation in MPI_Init is the single greatest consumer of time during startup. To-date, we have executed that operation as an ORTE stage gate that held the process until a startup message containing all required modex (and OOB contact info - see #3 below) info could be sent to it. Each process would send its data to the HNP's registry, which assembled and sent the message when all processes had reported in.

In addition, ORTE had taken responsibility for monitoring process status as it progressed through a series of "stage gates". The process reported its status at each gate, and ORTE would then send a "release" message once all procs had reported in.

The incoming changes revamp these procedures in three ways:

1. eliminating the ORTE stage gate system and cleanly delineating responsibility between the OMPI and ORTE layers for MPI init/finalize. The modex stage gate (STG1) has been replaced by a collective operation in the modex itself that performs an allgather on the required modex info. The allgather is implemented using the orte_grpcomm framework since the BTL's are not active at that point. At the moment, the grpcomm framework only has a "basic" component analogous to OMPI's "basic" coll framework - I would recommend that the MPI team create additional, more advanced components to improve performance of this step.

The other stage gates have been replaced by orte_grpcomm barrier functions. We tried to use MPI barriers instead (since the BTL's are active at that point), but - as we discussed on the telecon - these are not currently true barriers so the job would hang when we fell through while messages were still in process. Note that the grpcomm barrier doesn't actually resolve that problem, but Brian has pointed out that we are unlikely to ever see it violated. Again, you might want to spend a little time on an advanced barrier algorithm as the one in "basic" is very simplistic.

Summarizing this change: ORTE no longer tracks process state nor has direct responsibility for synchronizing jobs. This is now done via collective operations within the MPI layer, albeit using ORTE collective communication services. I -strongly- urge the MPI team to implement advanced collective algorithms to improve the performance of this critical procedure.


2. reducing the volume of data exchanged during modex. Data in the modex consisted of the process name, the name of the node where that process is located (expressed as a string), plus a string representation of all contact info. The nodename was required in order for the modex to determine if the process was local or not - in addition, some people like to have it to print pretty error messages when a connection failed.

The size of this data has been reduced in three ways:

(a) reducing the size of the process name itself. The process name consisted of two 32-bit fields for the jobid and vpid. This is far larger than any current system, or system likely to exist in the near future, can support. Accordingly, the default size of these fields has been reduced to 16-bits, which means you can have 32k procs in each of 32k jobs. Since the daemons must have a vpid, and we require one daemon/node, this also restricts the default configuration to 32k nodes.

To support any future "mega-clusters", a configuration option --enable-jumbo-apps has been added. This option increases the jobid and vpid field sizes to 32-bits. Someday, if necessary, someone can add yet another option to increase them to 64-bits, I suppose.

(b) replacing the string nodename with an integer nodeid. Since we have one daemon/node, the nodeid corresponds to the local daemon's vpid. This replaces an often lengthy string with only 2 (or at most 4) bytes, a substantial reduction.

(c) when the mca param requesting that nodenames be sent to support pretty error messages, a second mca param is now used to request FQDN - otherwise, the domain name is stripped (by default) from the message to save space. If someone wants to combine those into a single param somehow (perhaps with an argument?), they are welcome to do so - I didn't want to alter what people are already using.

While these may seem like small savings, they actually amount to a significant impact when aggregated across the entire modex operation. Since every proc must receive the modex data regardless of the collective used to send it, just reducing the size of the process name removes nearly 400MBytes of communication from a 32k proc job (admittedly, much of this comm may occur in parallel). So it does add up pretty quickly.


3. routing RML messages to reduce connections. The default messaging system remains point-to-point - i.e., each proc opens a socket to every proc it communicates with and sends its messages directly. A new option uses the orteds as routers - i.e., each proc only opens a single socket to its local orted. All messages are sent from the proc to the orted, which forwards the message to the orted on the node where the intended recipient proc is located - that orted then forwards the message to its local proc (the recipient). This greatly reduces the connection storm we have encountered during startup.

It also has the benefit of removing the sharing of every proc's OOB contact with every other proc. The orted routing tables are populated during launch since every orted gets a map of where every proc is being placed. Each proc, therefore, only needs to know the contact info for its local daemon, which is passed in via the environment when the proc is fork/exec'd by the daemon. This alone removes ~50 bytes/process of communication that was in the current STG1 startup message - so for our 32k proc job, this saves us roughly 32k*50 = 1.6MBytes sent to 32k procs = 51GBytes of messaging.

Note that you can use the new routing method by specifying -mca routed tree - if you so desire. This mode will become the default at some point in the future.


There are a few minor additional changes in the commit that I'll just note in passing:

* propagation of command line mca params to the orteds - fixes ticket #1073. See note there for details.

* requiring of "finalize" prior to "exit" for MPI procs - fixes ticket #1144. See note there for details.

* cleanup of some stale header files

This commit was SVN r16364.
2007-10-05 19:48:23 +00:00
Ralph Castain
f80ea093a2 Ensure that the orteds do not directly respond to USR1/2 signals. Those signals are trapped by mpirun and propagated from there - at most, the orteds are involved in the propagation process, but should never do anything on their own.
This commit was SVN r16098.
2007-09-12 14:32:31 +00:00
Sven Stork
5fd6c69019 - fix a problem showed up with the sun thread tests.
Remove unnecessary locks because functions that are calling this
  function proper lock/unlock the orted_comm_mutex. Therefore this 
  unlocks cause some imballance.

This commit was SVN r15630.
2007-07-26 11:30:27 +00:00