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.
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 The modex stage gate (STG1) has been replaced by a collective operation in the modex itself that performs an allgather on the required modex info.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.
static builds for OMPI components that required extra LIBS or LDFLAGS
(e.g., the openib BTL).
Fixes trac:1155.
This commit was SVN r16314.
The following SVN revision numbers were found above:
r15900 --> open-mpi/ompi@50941ec389
The following Trac tickets were found above:
Ticket 1155 --> https://svn.open-mpi.org/trac/ompi/ticket/1155
Check if an exclusion string (i.e. '-mca btl ^sm) was provided; if so OFUD just disables itself.
This commit was SVN r16307.
The following Trac tickets were found above:
Ticket 1154 --> https://svn.open-mpi.org/trac/ompi/ticket/1154
Each one of them has a field to store QP type, but this is redundant.
Store qp type only in one structure (the component one).
This commit was SVN r16272.
working with Automake 1.10. This is a new hack, which should be much
more flexible. The ras doesn't contain any Objective C, so remove the
hack entirely from that Makefile.am.
This commit was SVN r16269.
(Someday I'll learn to do this before committing)
This commit was SVN r16260.
The following SVN revision numbers were found above:
r16252 --> open-mpi/ompi@e10f476c87
and implementation. This has shown drastic performance benefit when
transferring Many files at roughly the same time.
I tested this for many different filem operations and everything was working
fine. Let me know if you have any problems with this functionality.
Some Notes:
- opal-checkpoint now has a 'quiet' flag to keep it from being too verbose.
- FileM RSH component is fully non-blocking.
- FileM RSH component has incomming connection throttling since by default
ssh only allows 10 concurrent scp connections to any single host. This
default can be adjusted via an MCA parameter.
{{{-mca filem_rsh_max_incomming 10}}}
- There is an MCA parameter for max outgoing connections, but it is currently
not implemented. If someone needs it then it should not be hard to implement.
{{{-mca filem_rsh_max_outgoing 10}}}
- Changed the FileM request structure so that it is a bit more explicit and
flexible.
- Moved the 'preload-binary' and 'preload-files' functionality into odls/base
allowing for code reuse in the 'process' and 'default' ODLS components.
- Fixed a bug in the process name resolution which broke the 'preload-*'
functionality due to GPR table structure changes.
- The FileM RSH component might be able to see even more speedup from using a
thread pool to operate on the work_pool structures, but that is for future
work.
- Added a 'opal-show-help' file to ODLS Base
This commit was SVN r16252.
performance characterization, and should not be used by anyone doing anything
else since it will not produce a globally consistent checkpoint in this mode.
This commit was SVN r16192.
is set to the cid of the communicator (unique id for each communicator).
Make sure each communicator have a group attached to it. The MPI_COMM_NULL
should have the MPI_GROUP_NULL as a group, in all circumstances.
This commit was SVN r16177.
executable has called MPI_Finalize(). It happens when removing the group
from each of the communicators, that MPI_COMM_NULL doesn't have a group.
Also fix the code from skipping over every other communicator when
freeing the groups.
This commit was SVN r16166.
Minor change to ompi_mca.m4 to move AC_CONFIG_FILES(framework/makefile) in autogen process (instead of configure process), where we still now the actual framework path (instead of guessing using $project/mca/$framework).
This have shown no side effects in our testing. Let us know if this breaks one of your components in some exotic context.
This commit was SVN r16146.
