* Clean up the DVM so it continues to run even when applications error out and we would ordinarily abort the daemons.
* Create a new errmgr component for the DVM to handle the differences.
* Cleanup the DVM state component.
* Add ORTE bindings directory and brief README
* Pass a local tool index around to match jobs.
* Pass the jobid on job completion.
* Fix initialization logic.
* Add framework for python wrapper.
* Fix terminate-with-non-zero-exit behavior so it properly terminates only the indicated procs, notifies orte-submit, and orte-dvm continues executing.
* Add some missing options to orte-dvm
* Fix a bug in -host processing that caused us to ignore the #slots designator. Add a new attribute to indicate "do not expand the DVM" when submitting job spawn requests.
* It actually makes no sense that we treat the termination of all children differently than terminating the children of a specific job - it only creates confusion over the difference in behavior. So terminate children the same way regardless.
Extend the cmd_line utility to easily allow layering of command line definitions
Catch up with ORTE interface change and make build more generic.
Disable "fixed dvm" logic for now.
Add another cmd_line function to merge a table of cmd line options with another one, reporting as errors any duplicate entries. Use this to allow orterun to reuse the orted_submit code
Fix the "fixed_dvm" logic by ensuring we reset num_new_daemons to zero. Also ensure that the nidmap is sent with the first job so the downstream daemons get the node info. Remove a duplicate cmd line entry in orterun.
Revise the DVM startup procedure to pass the nidmap only once, at the startup of the DVM. This reduces the overhead on each job launch and ensures that the nidmap doesn't get overwritten.
Add new commands to get_orted_comm_cmd_str().
Move ORTE command line options to orte_globals.[ch].
Catch up with extra orte_submit_init parameter.
Add example code.
Add documentation.
Bump version.
The nidmap and routing data must be updated prior to propagating the xcast or else the xcast will fail.
Fix the return code so it is something more expected when an error occurs. Ensure we get an error returned to us when we fail to launch for some reason. In this case, we will always get a launch_cb as we did indeed attempt to spawn it. The error code will be returned in the complete_cb.
Fix the return code from orte_submit_job - it was returning the tracker index instead of "success". Take advantage of ORTE's pretty-print capabilities to provide a nice error output explaining why we failed to launch. Ensure we always get a launch_cb when we fail to launch, but no complete_cb as the job never launched.
Extend the error reporting capability to job completion as well.
Add index parameter to orte_submit_job().
Add orte_job_cancel and implement ORTE_DAEMON_TERMINATE_JOB_CMD.
Factor out dvm termination.
Parse the terminate option at tool level.
Add error string for ORTE_ERR_JOB_CANCELLED.
Add some safeguards.
Cleanup and/of comments.
Enable the return.
Properly ORTE_DECLSPEC orte_submit_halt.
Add orte_submit_halt and orte_submit_cancel to interface.
Use the plm interface to terminate the job
This commit does two things. It removes checks for C99 required
headers (stdlib.h, string.h, signal.h, etc). Additionally it removes
definitions for required C99 types (intptr_t, int64_t, int32_t, etc).
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
Changing the client to leave its socket as blocking during the connect doesn't solve the problem by itself - you also have to introduce a sleep delay once the backlog is hit to avoid simply machine-gunning your way thru retries. This gets somewhat difficult to adjust as you don't want to unnecessarily prolong startup time.
We've solved this before by adding a listening thread that simply reaps accepts and shoves them into the event library for subsequent processing. This would resolve the problem, but meant yet another daemon-level thread. So I centralized the listening thread support and let multiple elements register listeners on it. Thus, each daemon now has a single listening thread that reaps accepts from multiple sources - for now, the orte/pmix server and the oob/usock support are using it. I'll add in the oob/tcp component later.
This still didn't fully resolve the SMP problem, especially on coprocessor cards (e.g., KNC). Removing the shared memory dstore support helped further improve the behavior - it looks like there is some kind of memory paging issue there that needs further understanding. Given that the shared memory support was about to be lost when I bring over the PMIx integration (until it is restored in that library), it seemed like a reasonable thing to just remove it at this point.
WHAT: Merge the PMIx branch into the devel repo, creating a new
OPAL “lmix” framework to abstract PMI support for all RTEs.
Replace the ORTE daemon-level collectives with a new PMIx
server and update the ORTE grpcomm framework to support
server-to-server collectives
WHY: We’ve had problems dealing with variations in PMI implementations,
and need to extend the existing PMI definitions to meet exascale
requirements.
WHEN: Mon, Aug 25
WHERE: https://github.com/rhc54/ompi-svn-mirror.git
Several community members have been working on a refactoring of the current PMI support within OMPI. Although the APIs are common, Slurm and Cray implement a different range of capabilities, and package them differently. For example, Cray provides an integrated PMI-1/2 library, while Slurm separates the two and requires the user to specify the one to be used at runtime. In addition, several bugs in the Slurm implementations have caused problems requiring extra coding.
All this has led to a slew of #if’s in the PMI code and bugs when the corner-case logic for one implementation accidentally traps the other. Extending this support to other implementations would have increased this complexity to an unacceptable level.
Accordingly, we have:
* created a new OPAL “pmix” framework to abstract the PMI support, with separate components for Cray, Slurm PMI-1, and Slurm PMI-2 implementations.
* Replaced the current ORTE grpcomm daemon-based collective operation with an integrated PMIx server, and updated the grpcomm APIs to provide more flexible, multi-algorithm support for collective operations. At this time, only the xcast and allgather operations are supported.
* Replaced the current global collective id with a signature based on the names of the participating procs. The allows an unlimited number of collectives to be executed by any group of processes, subject to the requirement that only one collective can be active at a time for a unique combination of procs. Note that a proc can be involved in any number of simultaneous collectives - it is the specific combination of procs that is subject to the constraint
* removed the prior OMPI/OPAL modex code
* added new macros for executing modex send/recv to simplify use of the new APIs. The send macros allow the caller to specify whether or not the BTL supports async modex operations - if so, then the non-blocking “fence” operation is used, if the active PMIx component supports it. Otherwise, the default is a full blocking modex exchange as we currently perform.
* retained the current flag that directs us to use a blocking fence operation, but only to retrieve data upon demand
This commit was SVN r32570.
http://www.open-mpi.org/community/lists/devel/2014/05/14822.php
Revamp the ORTE global data structures to reduce memory footprint and add new features. Add ability to control/set cpu frequency, though this can only be done if the sys admin has setup the system to support it (or you run as root).
This commit was SVN r31916.
Paul Hargrove pointed out that Stevens tells us that we should
FD_GETFL before FD_SETFL. And so we shall.
Make a new convenience function to do this (opal_fd_set_cloexec()),
just so that we don't have to litter this 2-step process throughout
the code.
Refs trac:4550
This commit was SVN r31513.
The following Trac tickets were found above:
Ticket 4550 --> https://svn.open-mpi.org/trac/ompi/ticket/4550
To accompany r31092 and r310924, also ensure to create a new process
group in the child right after the orted forks. Add trivial configury
to ensure that we have setpgid, and only do the setpgid/getpgid if we
have setpgid.
Without this commit, killing the entire process group can do
unexpected things (e.g., kill the orted, mpirun, and even mpirun's
parent!).
cmr=v1.7.5:reviewer=rhc
This commit was SVN r31132.
The following SVN revision numbers were found above:
r31092 --> open-mpi/ompi@99c9ecaed0
The following SVN revisions from the original message are invalid or
inconsistent and therefore were not cross-referenced:
r310924
The problem arises when a hostfile is used, and the user provides host names without specifying the slots= paramater. In these cases, we assign slots=1, but automatically allow oversubscription since that number isn't confirmed. We then provide a separate parameter by which the user can direct that we assign the number of slots based on the sensed hardware - e.g., by telling us to set the #slots equal to the #cores on each node. However, this has been set to "off" by default.
In order to make this a little less complex for the user, set the default such that we automatically set #slots equal to #cores (or #hwt's if use_hwthreads_as_cpus has been set) only for those cases where the user provides names in a hostfile but does not provide slot information.
Also cleanup some a couple of issues in the mapping/binding system:
* ensure we only override the binding directive if we are oversubscribed *and* overload is not allowed
* ensure that the MPI procs don't attempt to bind themselves if they are launched by an orted as any binding directive (no matter what it was) would have been serviced by the orted on launch
* minor cleanup to the warning message when oversubscribed and binding was requested
cmr=v1.7.5:reviewer=rhc:subject=update mapping/binding system
This commit was SVN r30909.
Reset topology usage for each node as we bind as multiple nodes may be linked to the same topology object. This will need to be revisited for scale as it does take some non-zero time to reset the usage each iteration. However, storing individual topology objects for every node consumes memory, so it's a tradeoff.
cmr=v1.7.4:reviewer=jsquyres:subject=Eliminate excessive binding/memory warnings
This commit was SVN r29978.
Provide some nice error messages if we fail to set the limits. Since the user had to specifically request we set the limit, treat failure as an error-out situation.
This commit was SVN r28288.
Features:
- Support for an override parameter file (openmpi-mca-param-override.conf).
Variable values in this file can not be overridden by any file or environment
value.
- Support for boolean, unsigned, and unsigned long long variables.
- Support for true/false values.
- Support for enumerations on integer variables.
- Support for MPIT scope, verbosity, and binding.
- Support for command line source.
- Support for setting variable source via the environment using
OMPI_MCA_SOURCE_<var name>=source (either command or file:filename)
- Cleaner API.
- Support for variable groups (equivalent to MPIT categories).
Notes:
- Variables must be created with a backing store (char **, int *, or bool *)
that must live at least as long as the variable.
- Creating a variable with the MCA_BASE_VAR_FLAG_SETTABLE enables the use of
mca_base_var_set_value() to change the value.
- String values are duplicated when the variable is registered. It is up to
the caller to free the original value if necessary. The new value will be
freed by the mca_base_var system and must not be freed by the user.
- Variables with constant scope may not be settable.
- Variable groups (and all associated variables) are deregistered when the
component is closed or the component repository item is freed. This
prevents a segmentation fault from accessing a variable after its component
is unloaded.
- After some discussion we decided we should remove the automatic registration
of component priority variables. Few component actually made use of this
feature.
- The enumerator interface was updated to be general enough to handle
future uses of the interface.
- The code to generate ompi_info output has been moved into the MCA variable
system. See mca_base_var_dump().
opal: update core and components to mca_base_var system
orte: update core and components to mca_base_var system
ompi: update core and components to mca_base_var system
This commit also modifies the rmaps framework. The following variables were
moved from ppr and lama: rmaps_base_pernode, rmaps_base_n_pernode,
rmaps_base_n_persocket. Both lama and ppr create synonyms for these variables.
This commit was SVN r28236.
It appears the problem was not with the command line parser but the rsh plm. I don't know why this problem was not occuring before the command line parser changes but it appears to be resolved now.
This commit was SVN r27527.
The following SVN revision numbers were found above:
r27451 --> open-mpi/ompi@d59034e6ef
r27456 --> open-mpi/ompi@ecdbf34937
readable by normal human beings, vs. having a bitmap of physical
PU's). Use the new hwloc base prettyprint functions to generate the
output.
This commit was SVN r26533.
* Remove paffinity, maffinity, and carto frameworks -- they've been
wholly replaced by hwloc.
* Move ompi_mpi_init() affinity-setting/checking code down to ORTE.
* Update sm, smcuda, wv, and openib components to no longer use carto.
Instead, use hwloc data. There are still optimizations possible in
the sm/smcuda BTLs (i.e., making multiple mpools). Also, the old
carto-based code found out how many NUMA nodes were ''available''
-- not how many were used ''in this job''. The new hwloc-using
code computes the same value -- it was not updated to calculate how
many NUMA nodes are used ''by this job.''
* Note that I cannot compile the smcuda and wv BTLs -- I ''think''
they're right, but they need to be verified by their owners.
* The openib component now does a bunch of stuff to figure out where
"near" OpenFabrics devices are. '''THIS IS A CHANGE IN DEFAULT
BEHAVIOR!!''' and still needs to be verified by OpenFabrics vendors
(I do not have a NUMA machine with an OpenFabrics device that is a
non-uniform distance from multiple different NUMA nodes).
* Completely rewrite the OMPI_Affinity_str() routine from the
"affinity" mpiext extension. This extension now understands
hyperthreads; the output format of it has changed a bit to reflect
this new information.
* Bunches of minor changes around the code base to update names/types
from maffinity/paffinity-based names to hwloc-based names.
* Add some helper functions into the hwloc base, mainly having to do
with the fact that we have the hwloc data reporting ''all''
topology information, but sometimes you really only want the
(online | available) data.
This commit was SVN r26391.
Roll in the ORTE state machine. Remove last traces of opal_sos. Remove UTK epoch code.
Please see the various emails about the state machine change for details. I'll send something out later with more info on the new arch.
This commit was SVN r26242.
https://svn.open-mpi.org/trac/ompi/wiki/ProcessPlacement
The wiki page is incomplete at the moment, but I hope to complete it over the next few days. I will provide updates on the devel list. As the wiki page states, the default and most commonly used options remain unchanged (except as noted below). New, esoteric and complex options have been added, but unless you are a true masochist, you are unlikely to use many of them beyond perhaps an initial curiosity-motivated experimentation.
In a nutshell, this commit revamps the map/rank/bind procedure to take into account topology info on the compute nodes. I have, for the most part, preserved the default behaviors, with three notable exceptions:
1. I have at long last bowed my head in submission to the system admin's of managed clusters. For years, they have complained about our default of allowing users to oversubscribe nodes - i.e., to run more processes on a node than allocated slots. Accordingly, I have modified the default behavior: if you are running off of hostfile/dash-host allocated nodes, then the default is to allow oversubscription. If you are running off of RM-allocated nodes, then the default is to NOT allow oversubscription. Flags to override these behaviors are provided, so this only affects the default behavior.
2. both cpus/rank and stride have been removed. The latter was demanded by those who didn't understand the purpose behind it - and I agreed as the users who requested it are no longer using it. The former was removed temporarily pending implementation.
3. vm launch is now the sole method for starting OMPI. It was just too darned hard to maintain multiple launch procedures - maybe someday, provided someone can demonstrate a reason to do so.
As Jeff stated, it is impossible to fully test a change of this size. I have tested it on Linux and Mac, covering all the default and simple options, singletons, and comm_spawn. That said, I'm sure others will find problems, so I'll be watching MTT results until this stabilizes.
This commit was SVN r25476.
Fix a termination issue that caused procs local to mpirun to not be killed if they weren't calling into the library. Thanks to Terry Dontje for spending countless hours chasing his tail on this one! :-(
This commit was SVN r25285.
extravaganza.
= Short version =
This commit does several things, but the short version is that it
re-orients the error message creation of the ODLS default module to
generate error strings in the child process for errors that occur
after the fork but before the exec (such errors are ''usually''
related to paffinity). A show_help string is rendered in the child
and then IPC'ed up to the parent, who displays the string through
normal ORTE show_help aggregation mechanisms. We also broke up the
ginormous paffinity-setting logic into a few separate functions, both
to help us understand the code, and hopefully to ease future
maintenance.
The logic for the ODLS default binding should not have changed -- this
is mainly a code reshuffle and improvement on error reporting.
= Rationale =
The reasoning for this commit is complex. As mentioned above, it's
the first step in some paffinity cleanup. Here's the line of dominoes
that must fall (in this order):
1. Add hwloc paffinity component (already done).
1. While testing hwloc, we discovered that the error reporting from
the ODLS default module was abysmal. So we fixed it.
1. Further, we reorganized the code in the odsl_default_module.c a bit
to help our understanding of it.
1. We also discovered a few bugs in the original ODLS default module
logic that existed before this code shuffle; separate tickets
will be filed to fix them.
1. Next up will be some improvements to paffinity / odls default to
make the act of binding to a core ensure to bind to ''all''
hardware threads contained in that core (similar for sockets:
binding to a socket will bind to ''all'' hardware threads in that
socket).
1. Next will be improvements to paffinity to expose binding to
hardware threads through the paffinity framework API.
1. Finally, we'll expose these binding controls to the user (e.g.,
through mpirun command line arguments, MCA parameters, etc.).
This commit represents the first few bullets; the last 4 bullets are
being worked on right now, but there is no definite timeline for
completion.
= Miscelaneous =
A few points worth mentioning:
* We have tested this new code a bunch; we're pretty sure it behaves
just like the trunk -- but with better / more precise error
reporting. More testing is needed on a wider array of platforms,
however.
* A big comment at the top of odls_default_module.c explains the
(new) general scheme for the error reporting.
* The error reporting in the parent process is now really dumb;
almost all the intelligence about creating error messages is in the
child.
* The show_help file was renamed to be more consistent with other
help files (help-odls-default.txt -> help-orte-odls-default.txt)
* Removed the use of sched_yield() because of recent changes in the
Linux 2.6.3x kernels. We already had an #else clause for
select()'ing for 1us if we didn't have sched_yield() -- that is now
the only code path. This is not a performance-critical section of
the code, so this shouldn't be controversial.
* Replaced the macro-based error reporting with function-based
reporting. It's a bit more bulky, but it helped us understand the
code and saved us multiple times with compile-time parameter
checking, etc.
* Cleaned up the use of several show_help messages to ensure that
they mapped to real messages in help*.txt files.
This commit was SVN r23652.
* Remove OPAL_ERR_PAFFINITY_NOT_SUPPORTED; fit it into the generic
OPAL_ERR_NOT_SUPPORTED case.
* When odls_default detects that processor affinity is not supported,
it prints a specific message about it, and then it suppressed a
generic HNP help message that would normally follow it (i.e., it's
easier to have the "processor affinity is not supported" show_help
message last).
* Use some symbolic names in odls_default instead of fixed int's,
just for slight readability improvements in the code.
* Introduce orte_show_help_suppress(), which gives the ability to
suppress any future showings of any arbitrary show_help() message.
This is useful if you display message X and want to suppress
message Y. This suppression *only* works in environments where
orte_show_help() does coalescing.
This commit was SVN r23249.
* If < 0, it's an OPAL_ERR_* value
* If >= 0, it's the actual output value of the function
This is problematic for the OPAL_SOS stuff. This commit changes those
functions to always return OPAL_* statuses and send the output value
back through output parameters (like 95% of the rest of the code
base). This avoids the confusion with OPAL_SOS stuff and makes
paffinity work again (e.g., mpirun --bind-to-core ...).
I updated all paffinitiy modules for the new function signatures, and
bumped the paffinity API version up to 2.0.1. I don't think the
version change will matter, though, because we'll be introducing
support for hardware threads soon, which will either bump the
paffinity version again or we'll replace paffinity with
a new framework.
This commit was SVN r23197.
