1
1

20 Коммитов

Автор SHA1 Сообщение Дата
Ralph Castain
3cffa6b549 Inserts some debugging output for use by Tim to help debug a startup problem. Should be transparent to everyone else.
This commit was SVN r2453.
2004-09-02 18:39:42 +00:00
Jeff Squyres
7f83536f7a Fix a bunch of RUI's and ensure that all the Right functions are
invoked from the various finalize functions so that we don't have
resource leaks

This commit was SVN r2374.
2004-08-29 09:05:14 +00:00
Tim Woodall
16d250b376 - integration of gpr/ns/oob w/ mpirun2
This commit was SVN r2344.
2004-08-28 01:15:19 +00:00
Tim Woodall
3d754c4941 - modifications to oob to support bringing up oob before anything else (ns)
- changed ns/gpr to register callback in component init (and ignore errors if not supported)

This commit was SVN r2303.
2004-08-25 17:39:08 +00:00
Ralph Castain
b2868d0158 Begin knitting the openmpi program using the name server, registry, etc. Changes include:
- modified rte_init to utilize name server, interface to session directory system
- added fields to system_info and proc_info
- created a utility for converting process into daemon
- modified openmpi to interface to various functions, spawn universe, record contact info, etc. Program almost complete.
- began building ompid daemon

This commit was SVN r2270.
2004-08-23 21:39:22 +00:00
Tim Woodall
e05e7ca942 gpr proxy i/f (put) seems to be working w/ tcp oob!
This commit was SVN r2235.
2004-08-19 19:34:37 +00:00
David Daniel
927ef83454 This is a large check-in related to switching from
os/atomic.h  -->  include/sys/atomic.h

WARNING:  There are almost certainly some bugs introduced here, but I
believe that using this system will get us to a stable and portable
library faster in the long run.

Other changes:

threads/mutex

  Reorganized to use pthreads or asm atomic operations as available.
  Untested Windows implementation added using InterlockedExchange() funcion.

threads/thread

  Added an untested Windows implementation

other places

  Updates to include the "right" header files or to use
  ompi_atomic_add_int() rather that fetchNadd() etc.

This commit was SVN r2221.
2004-08-18 23:24:27 +00:00
Tim Woodall
195f83a769 call back into selected NS module to allow registration of OOB callbacks
This commit was SVN r2204.
2004-08-18 15:53:13 +00:00
Brian Barrett
7f3f72f1c7 * split the PCM into pcm and pcmclient to better match to the multi-pcm case
* add env pcmclient, which builds client data out of environment variables
* improved rsh infrastructure a little bit - almost able to get to calling
  fork()
* Removed instance information for llm - it should be burried in the pcm,
  not exposed


*** YOU MUST RUN AUTOGEN / CONFIGURE / etc AFTER UPDATING ***

This commit was SVN r2191.
2004-08-17 22:24:17 +00:00
Brian Barrett
d5133eddec * clean up the Doxygen comments in the runtime and llm classes
* recalarify previous clarification on ompi_list_get_size()

This commit was SVN r2190.
2004-08-17 22:14:47 +00:00
Brian Barrett
5540dc37bc * Change from ompi_list_t to ompi_list_t* in the schedule and allocation
structures to make it easier to swap around lists when doing process ->
  resource mapping
* Fix spawn interface to take an ompi_list_t* instead of an ompi_list_t
  since you can't pass an ompi_list_t by value
* Change allocate_resource to return an ompi_list_t* instead of having
  an ompi_list_t** as an argument, since it's a bit cleaner and makes
  who should call OBJ_NEW much more clear
* Clean up deallocation in error cases for the llm_base_allocate function
* Update test case for llm to not depend on current environment for
  correctness

This commit was SVN r2126.
2004-08-13 19:39:06 +00:00
Tim Prins
6c40301350 Added #define _GNU_SOURCE so we can compile on all linux systems.
This commit was SVN r2007.
2004-08-10 18:29:52 +00:00
Brian Barrett
7932c850c1 * Move the proc_info and session_dir initialization into the ompi_rte_init
code
* clean up comments in the session and proc code so that it's readable in
  80 columns
* Try to get cleanup correct

This commit was SVN r2003.
2004-08-10 16:05:22 +00:00
Tim Prins
457a4b632c Some more changes to the oob:
All the user interface functions are now in mca/oob/base/base.h
Anyone who uses the oob should just include this file.

All component related functions have been moved to mca/oob/oob.h

The reason for this change was to make the user interface more clear.

This commit was SVN r1884.
2004-08-04 23:42:51 +00:00
Jeff Squyres
6c3159ec0b Add some printf's to know where various init functions fail, until we
have show_help()

This commit was SVN r1670.
2004-07-13 12:35:43 +00:00
Jeff Squyres
4c14a1cd7e Refinement of the first cut of the ns framework.
- added in selection logic for ns
- made all types consistent in ns
- made replica component match types and base selection logic; it only
  allows itself to be selected in the seed
- copied replica component to "proxy" and altered the selection logic;
  it only allows itself to be selected when *not* in the seed.  All
  of the body functions in proxy need to be implemented; this is only
  a skeleton for now so that MPI programs have a ns component to
  select (because they're not the seed) and can therefore run
  properly.
- added mca_ns_base_open/close to ompi_rte_init/finalize

This commit was SVN r1621.
2004-07-11 04:34:47 +00:00
Tim Woodall
165d4f8841 merge of oob/rte changes from rte branch
This commit was SVN r1527.
2004-07-01 14:49:54 +00:00
Jeff Squyres
e88c37e1f8 Fix logic in determining whether all modules support multiple user
threads or not.

This commit was SVN r1361.
2004-06-17 15:08:07 +00:00
Jeff Squyres
1b542d409b Remove all remnants of the component framework "registry".
Registry is dead; long live gpr!

This commit was SVN r1359.
2004-06-17 13:53:10 +00:00
David Daniel
2f96ba71fe renaming files
This commit was SVN r1192.
2004-06-07 15:40:19 +00:00