1
1
Граф коммитов

2089 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
a14ff905f8 Update the convertor.c file. This commit was supposed to go together with
r12486 but somehow I miss it.
Update the pack and unpack functions for contigusous datatypes to minimize their
impact on the performance. Keep them as condensed as possible.

This commit was SVN r12513.

The following SVN revision numbers were found above:
  r12486 --> open-mpi/ompi@8746369338
2006-11-09 17:26:55 +00:00
George Bosilca
a82ce427e4 Update the number of reduce algorithms available.
This commit was SVN r12503.
2006-11-08 22:20:34 +00:00
George Bosilca
0dcf0097db They are supposed to be ints not size_t.
This commit was SVN r12497.
2006-11-08 17:05:15 +00:00
George Bosilca
eab1776e9a Explicit casts for our friendly Windows environment...
This commit was SVN r12496.
2006-11-08 17:02:46 +00:00
George Bosilca
f9f1087e7f Be nice and let everybody access this variable.
This commit was SVN r12495.
2006-11-08 16:56:17 +00:00
George Bosilca
0914892044 Small cleanups, some explicit casts.
This commit was SVN r12494.
2006-11-08 16:54:03 +00:00
George Bosilca
74d3946342 Remove the call to set_args. This is only required for the MPI level,
because there we have to be able to return to the user the description
of the data.

This commit was SVN r12493.
2006-11-08 16:52:48 +00:00
George Bosilca
b53a313904 These functions dont have to be exported. The outside world
access them only via the pointer in the convertor structure.

This commit was SVN r12492.
2006-11-08 16:49:34 +00:00
George Bosilca
8746369338 This commit was supposed to go with the r12485.
This commit was SVN r12486.

The following SVN revision numbers were found above:
  r12485 --> open-mpi/ompi@b4e9199651
2006-11-08 06:21:15 +00:00
George Bosilca
b4e9199651 Minimize the impact on contiguous and predefined data-types. Most of the
usefull things are now done without the call to any external function,
directly in the macros in this file.
Create a new flag for the convertor to help figure out when we have
to do anything special for pack/unpack. A convertor with the flag
CONVERTOR_NO_OP is a basic convertor, designed for contiguous types.

This commit was SVN r12485.
2006-11-08 06:16:47 +00:00
George Bosilca
915d748d72 Initialize the convertor on _START not on _INIT. This allow us to
set it up before the match when we know the peer, saving some
time on the critical path. If the receive is ANY_SOURCE then
we initialize the convertor on _MATCHED. Anyway, we will set it
up only once per receive.

This commit was SVN r12484.
2006-11-08 05:42:29 +00:00
George Bosilca
eb45a5e402 Move things around a little bit. Mainly fields from the send and receive
request in the base request. Rearrange the fields to keep the data
together. Remove some useless tests.

This commit was SVN r12482.
2006-11-08 04:58:23 +00:00
George Bosilca
63462331c9 Reduce the number of branches. Keep the fast path as short as possible.
Remove some useless error checking. Add OPAL_UNLIKELY directives.

This commit was SVN r12477.
2006-11-07 23:59:32 +00:00
George Bosilca
f3de2e1a82 Keep the fast path as short as possible.
This commit was SVN r12476.
2006-11-07 23:56:32 +00:00
Jeff Squyres
31e8f510fe After review with Rolf, decided to be a bit more safe and instead of
assigning status._cancelled, call MPI_STATUS_SET_CANCELLED.

This commit was SVN r12471.
2006-11-07 20:49:31 +00:00
Jeff Squyres
427c20af0d Use a new algorithm for allgatherv. The old algorithm essentially did
N gatherv's:

  for (i = 0 ... size)
    MPI_Gatherv(..., root = i, ...)

The new algorithm simply does (effectively):

  MPI_Gatherv(..., root = 0, ...)
  MPI_Bcast(..., root = 0, ...)

This commit was SVN r12469.
2006-11-07 18:07:55 +00:00
Jeff Squyres
25dab9700f * Add some error checking to GET_ELEMENTS and STATUS_SET_CANCELLED
* Move the check for STATUS_IGNORE in STATUS_SET_ELEMENTS up into the
   error checking block

This commit was SVN r12460.
2006-11-07 02:47:35 +00:00
Galen Shipman
55db17b37c don't try to use a dead btl..
This commit was SVN r12456.
2006-11-06 23:25:24 +00:00
George Bosilca
f81a599326 Don't do the safe guard for contiguous memory layouts.
This commit was SVN r12455.
2006-11-06 23:15:24 +00:00
George Bosilca
108ea4dbe9 When the MX MTL complete a request, force a return from the progress function.
Decrease the latency by about 0.3 microseconds.

This commit was SVN r12454.
2006-11-06 23:13:07 +00:00
George Bosilca
3d0df2cf29 Allow the MX BTL to finish the small sends quicker. Once the mx_isend is posted if
the message size is less than 4K do a check for the message completion and if any
call the callback.

This commit was SVN r12453.
2006-11-06 23:12:01 +00:00
Galen Shipman
eef37430a7 failing already failed for ACK timeout..
This commit was SVN r12452.
2006-11-06 22:09:39 +00:00
Galen Shipman
813e7faea8 more fixes for failover.. and yet still more to come..
This commit was SVN r12450.
2006-11-06 21:27:17 +00:00
Jeff Squyres
6e1729cb93 Fixes trac:580
* Add MPI::Status methods Set_elements() and Set_cancelled()
 * Added a bunch of comments in various places in the MPI C++ bindings
   implementatio just to explain what's going on (because C++ can hide
   a lot from you)
 * Insert C++ callbacks for the MPI_Grequest callback functions
   registered by MPI::Grequest::Start().  These callbacks keep a
   little meta-data (created by Grequest::Start()) that allow the
   proper callback signatures from C (i.e., from ompi_grequest_<foo>()
   in libmpi.a -- C code), translate arguments as required, and then
   invoke the callbacks with proper C++ signatures (i.e., call
   user-defined callbacks with C++ function signatures).

This commit was SVN r12446.

The following Trac tickets were found above:
  Ticket 580 --> https://svn.open-mpi.org/trac/ompi/ticket/580
2006-11-06 18:42:00 +00:00
Gleb Natapov
b4fd2d7d50 Fix warnings from progress thread patch.
This commit was SVN r12434.
2006-11-06 12:34:56 +00:00
Gleb Natapov
82f7c0dd69 Fix regression from v1.1.
1) make the code do what comment says
2) if memory is prepinned don't send multiple PUT messages.

This commit was SVN r12433.
2006-11-06 12:00:17 +00:00
Ralph Castain
c77f6c605e Update timing reports:
1. Remove timing of xcast from mpi_init

2. Add timing report from oob_xcast on how long it took to send the message

This commit was SVN r12428.
2006-11-03 18:55:05 +00:00
Ralph Castain
60e27c77e7 Add some additional timing reporting:
1. Added reporting points around the xcasts in MPI_Init. Note that these times will include time spent waiting for a trigger to fire, which is why the times between stage gates did NOT include these times initially. The inter-stage-gate times still do NOT include the xcast time - the xcast time is reported separately.

2. Added the process vpid on the MPI_Init timing reports for clarity.

3. Added a report from the xcast function on the HNP that outputs the number of bytes in the message being sent to the processes.

This commit was SVN r12422.
2006-11-03 16:04:40 +00:00
Jeff Squyres
bc93583184 Fix the SIZEOF implementation for complex's. We want to output the
size of the complex type as determined by configure; not the size of
the next larger complex type (i.e., a complex*N is 2 real*(N/2)'s, not
2 real*N's).

This commit was SVN r12421.
2006-11-03 15:52:46 +00:00
Galen Shipman
f7c554df65 Try to failover when we get an async error from the lower layer (BTL)..
This commit was SVN r12420.
2006-11-03 15:40:26 +00:00
George Bosilca
8529238d93 Add 2 more algorithms to the dynamic list.
This commit was SVN r12415.
2006-11-02 19:19:08 +00:00
George Bosilca
110d07b7d3 Small optimization or zero length messages.
This commit was SVN r12414.
2006-11-02 19:10:28 +00:00
Pavel Shamis
566667ac61 Adding progress thread support to OpenIB BTL.
Reviewed by Gleb.

This commit was SVN r12411.
2006-11-02 16:15:21 +00:00
George Bosilca
284e286011 Correctly compute the local size using the count argument not
the convertor count field which is not yet initialized.

This commit was SVN r12405.
2006-11-02 06:44:51 +00:00
Jeff Squyres
431f940a52 Fixes trac:496
* Add some more error checking to GREQUEST_START
 * Move the error checking in GREQUEST_COMPLETE up to inside the
   MPI_PARAM_CHECK block, where it belongs
 * Invoke the gen request query_fn in all the Right spots (per MPI-2:8.2)
 * Distinguish between grequests created from C and Fortran
 * Use the OBJ system to reference count to release the grequest at
   the Right time and invoke the grequest free_fn properly (see
   lengthy comment in grequest.c above the destructor)
 * Have ompi_grequest_complete() call ompi_request_complete() rather
   than [poorly] copy the contents of ompi_request_complete()
 * Fix Fortran function callback pointer typedefs to use proper
   Fortran types
 * Edit ompi_request_test* and ompi_request_wait* to properly handle
   generalized requests.  This adds an "if" statement in the critical
   path for all the back-end test* and wait* functions :-(,
   but fortunately George took out two "if" statements from the
   critical path last week.  So we're still ahead.  :-)
 * Move ompi_request_test() out of request.h and into request.c (all
   other test* and wait* functions were already in the .c file -- and
   ompi_request_test() was too long to be statically inlined anyway)

This commit was SVN r12402.

The following Trac tickets were found above:
  Ticket 496 --> https://svn.open-mpi.org/trac/ompi/ticket/496
2006-11-02 03:34:53 +00:00
George Bosilca
ea91fd3bdb Small optimizations and some cleanups.
This commit was SVN r12400.
2006-11-02 00:19:03 +00:00
George Bosilca
994bfce7e8 On heterogeneous environment check for the hetero mask before deciding if there is or not
anything to do for packing and unpacking of a particular data-type.

This commit was SVN r12399.
2006-11-01 23:57:37 +00:00
George Bosilca
d89a52f8e7 If there is no data to send or receive don't do anything useless.
This commit was SVN r12398.
2006-11-01 23:56:34 +00:00
George Bosilca
dbec514b0f Optimize the generation of the match_bits and the mask.
This commit was SVN r12396.
2006-11-01 23:19:20 +00:00
George Bosilca
23e783a487 Initialize the length before using it.
This commit was SVN r12394.
2006-11-01 22:07:21 +00:00
Jeff Squyres
2f936c4597 Fixes trac:549
Had group discussion about this on the weekly call.  The decision was
that we should pass the real error code to the back-end exception
handler because it's pretty useless to pass MPI_ERR_IN_STATUS to the
back-end exception handler (because exception handlers don't have
access to the request or the status - this has potential issues for
fault tolerance kinds of scenarios).  So in TESTALL, TESTSOME,
WAITALL, and WAITSOME, we examine the error code and if it's not
MPI_SUCCESS, return MPI_ERR_IN_STATUS.

This commit was SVN r12389.

The following Trac tickets were found above:
  Ticket 549 --> https://svn.open-mpi.org/trac/ompi/ticket/549
2006-11-01 19:31:43 +00:00
Gleb Natapov
4c784b6403 As Andrew Friedley pointed, my previous patch may cause deadlock if
mca_btl_openib_endpoint_connect_eager_rdma() is called recursively. He also
noticed that orte_pointer_array_add() can't fail because we allocate max number
of elements at init time. So just remove error handling and locking. No locking
 - no deadlocks.

This commit was SVN r12388.
2006-11-01 15:53:33 +00:00
Gleb Natapov
3bf31fe4a3 Correctly determine the first element on the list. opal_list_get_prev() never
returns NULL, it should be compared with opal_list_get_end() instead.

This commit was SVN r12387.
2006-11-01 13:44:47 +00:00
Gleb Natapov
b5714d698a Fix compilation with GM version smaller than 2.0. Fix compilation warnings.
This commit was SVN r12386.
2006-11-01 10:26:15 +00:00
Gleb Natapov
aac695a51f eager_rdma_buffers update is not atomic. A buffer is added to the array and if
something is going wrong down in the code it is removed from the array. So add
mutex to prevent concurrent access to the array from different threads.

This commit was SVN r12385.
2006-11-01 07:27:32 +00:00
Ralph Castain
9204747930 Add timing info to comm_spawn - timing collected and reported when OMPI_MCA_ompi_timing = 1 (or something other than zero).
This commit was SVN r12381.
2006-10-31 23:32:39 +00:00
George Bosilca
a711a58410 Add a function to handle the MPI_Status_set_elements correctly. The current
implementation was just wrong !!!

This commit was SVN r12380.
2006-10-31 23:02:42 +00:00
Ralph Castain
17c71f8d2a Fix ticket #545
Setup subscriptions to correctly return the MPI_APPNUM attribute.

Fix an unreported bug that was found. The universe size was incorrectly defined in the attributes code. As coded, it looked for size_t values and based its size computation on those numbers. Unfortunately, the node_slots value had been changed to an orte_std_cntr_t awhile back! So the universe size was never updated.

Update the hello_nodename test to check for MPI_APPNUM.

Add a definition to ns_types for ORTE_PROC_MY_NAME - just a shortcut for orte_process_info.my_name. Brought over from ORTE 2.0 as it will be used extensively there.

This commit was SVN r12377.
2006-10-31 21:29:07 +00:00
Andrew Friedley
48c5117476 Fix some signedness warnings on threaded builds introduced by r12369
This commit was SVN r12376.

The following SVN revision numbers were found above:
  r12369 --> open-mpi/ompi@d7375ec102
2006-10-31 17:29:25 +00:00
Gleb Natapov
d7375ec102 Fix deadlock reported by Andrew Friedley:
What's happening is that we're holding openib_btl->eager_rdma_lock when
we call mca_btl_openib_endpoint_send_eager_rdma() on
btl_openib_endpoint.c:1227.  This in turn calls
mca_btl_openib_endpoint_send() on line 1179.  Then, if the endpoint
state isn't MCA_BTL_IB_CONNECTED or MCA_BTL_IB_FAILED, we call
opal_progress(), where we eventually try to lock
openib_btl->eager_rdma_lock at btl_openib_component.c:997.

The fix removes this lock altogether. Instead we atomically set local RDMA
pointer to prevent other threads to create rdma buffer for the same endpoint.
And we increment eager_rdma_buffers_count atomically thus polling thread doesn't
need lock around it.

This commit was SVN r12369.
2006-10-31 09:54:52 +00:00