1
1
Граф коммитов

237 Коммитов

Автор SHA1 Сообщение Дата
George Bosilca
c2c6a1b37e Correctly compute the number of elements in a segment.
For broadcast send the correct size for all intermediary nodes.

This commit was SVN r12552.
2006-11-10 23:04:50 +00:00
George Bosilca
7102147b9f Correctly detect when the specified algorithm is out of range. In
this case we reset it to zero.

This commit was SVN r12551.
2006-11-10 21:47:07 +00:00
George Bosilca
af68171253 Use the macro to compute the number of elements in a segment in both
bcast and reduce and update the default values for the variables
as required by the comment in the coll_tuned.h file.

This commit was SVN r12546.
2006-11-10 20:04:08 +00:00
George Bosilca
476b922074 Updates & upgrades:
- consistent arguments checking (not allowing to select an algorithm which
     is not available)
 - consistent way of computing the segcount (number of datatypes by segment).
 - small cleanups.
 - more informative debugging messages.

This commit was SVN r12545.
2006-11-10 19:54:09 +00:00
George Bosilca
77ef979457 New architecture for broadcast. A generic broadcast working on a tree
description. Most of the bcast algorithms can be completed using this
generic function once we create the tree structure. Add all kind of
trees.

There are 2 versions of the generic bcast function. One using overlapping
between receives (for intermediary nodes) and then blocking sends to all
childs and another where all sends are non blocking. I still have to
figure out which one give the smallest overhead.

This commit was SVN r12530.
2006-11-10 05:53:50 +00:00
George Bosilca
1d80f685b5 Remove one compiler warning.
This commit was SVN r12520.
2006-11-09 20:08:43 +00:00
George Bosilca
73eec4bfef Show the MCA parameter coll_base_verbose only if Open MPI is compiled in
debug mode. Otherwise there is no debug anyway ...

This commit was SVN r12516.
2006-11-09 19:02:32 +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
0914892044 Small cleanups, some explicit casts.
This commit was SVN r12494.
2006-11-08 16:54:03 +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
393657ee26 Initialize the sndbuf in all cases. Do not forget to initialize the
tree used in each of the broadcast functions.

This commit was SVN r12332.
2006-10-27 00:13:33 +00:00
George Bosilca
ba3c247f2a Big collective commit. I lightly test it, but I think it should be quite stable. Anyway,
the default decision functions (for broadcast, reduce and barrier) are based on a
high performance network (not TCP). It should give good performance (really good) for
any network having the following caracteristics: small latency (5 microseconds) and good
bandwidth (more than 1Gb/s).
+ Cleanup of the reduce algorithms, plus 2 new algorithms (binary and binomial). Now most
  of the reduce algorithms use a generic tree based function for completing the reduce.
+ Added macros for computing the trees (they are used for bcast and reduce right now).
+ Allow the usage of all 5 topologies.
+ Jelena's implementation of a binary tree that can be used for non commutative operations.
  Right now only the tree building function is there, it will get activated soon.
+ Some others minor cleanups.

This commit was SVN r12326.
2006-10-26 22:53:05 +00:00
George Bosilca
99631ccf66 Cleanups.
This commit was SVN r12272.
2006-10-23 22:29:17 +00:00
George Bosilca
d7d3f9e486 Tuned collectives works only for at least 2 processes. We have the self module
for the other cases.

This commit was SVN r12271.
2006-10-23 22:28:56 +00:00
George Bosilca
b848a5ad06 Remove all ompi_coll_chain_t references.
This commit was SVN r12269.
2006-10-23 21:47:50 +00:00
George Bosilca
39cd8d3d17 One to rule them all. We only need one topology information: a tree. How we
build it it's hat make the difference.

This commit was SVN r12268.
2006-10-23 21:46:30 +00:00
George Bosilca
9cf3040e5f Allocate enough memory for the reduce operation when MPI_IN_PLACE is specified.
This commit was SVN r12260.
2006-10-23 17:51:36 +00:00
George Bosilca
6b697ad3dd If the operation is not commutative then force the basic reducve algorithm. The others
cannot be used for non commutative operations ... yet ...

This commit was SVN r12241.
2006-10-20 22:11:44 +00:00
George Bosilca
a7b6078b73 No more segfault. Still some wrong data around ...
This commit was SVN r12238.
2006-10-20 20:17:34 +00:00
George Bosilca
02759cf515 Update the reduce chain collective.
This commit was SVN r12237.
2006-10-20 19:47:52 +00:00
George Bosilca
06563b5dec Last set of explicit conversions. We are now close to the zero warnings on
all platforms. The only exceptions (and I will not deal with them
anytime soon) are on Windows:
- the write functions which require the length to be an int when it's
  a size_t on all UNIX variants.
- all iovec manipulation functions where the iov_len is again an int
  when it's a size_t on most of the UNIXes.
As these only happens on Windows, so I think we're set for now :)

This commit was SVN r12215.
2006-10-20 03:57:44 +00:00
George Bosilca
caefd6d0ee Do not leak memory. Allocate the intermediary buffer only when we really need it
(not leafs) and release on the same way.

This commit was SVN r12200.
2006-10-19 22:20:33 +00:00
George Bosilca
26b33ec2d7 If there is just one node, we don't need a decision function, just do the copy
and return.

This commit was SVN r12199.
2006-10-19 22:19:36 +00:00
George Bosilca
3eb2f90ceb For the recurvise doubling correctly compute the closest power of 2 number of
nodes.

This commit was SVN r12191.
2006-10-19 17:14:57 +00:00
George Bosilca
041fcb8d18 Update the barrier decision function.
This commit was SVN r12190.
2006-10-19 17:14:01 +00:00
George Bosilca
c9da782804 Keep only one function to get the size of a datatype.
This commit was SVN r12170.
2006-10-18 17:33:01 +00:00
George Bosilca
21ade43b96 Remove a non reacheable statement.
This commit was SVN r12166.
2006-10-18 16:43:55 +00:00
George Bosilca
be27ee6fa0 Correct the bcast problem where we always did a bcast with segzise of 0.
Activate the reduce decision function.
Others small updates (mostly TAB to spaces).

This commit was SVN r12161.
2006-10-18 02:00:46 +00:00
George Bosilca
8852c00c36 Look like a big commit but in fact it address only one issue. The way we're working with
size and diplacement of data-type. After this patch all data can contain size_t bytes
and the displacements are defined as ptrdiff_t. All of the files I was able to compile
have been modified to match this requirement.

This commit was SVN r12146.
2006-10-17 20:20:58 +00:00
Jeff Squyres
a8e9fa09da Fix some compiler warnings introduced in r11619. I checked with
George: ompi_ddt_type_size() returns a signed int only because of the
MPI spec; it will never return a negative value.  So casting the
return value out of it to a (uint32_t) is safe, and makes the
comparisons be between two unsigned values. 

This commit was SVN r11639.

The following SVN revision numbers were found above:
  r11619 --> open-mpi/ompi@8667648a1b
2006-09-13 16:42:31 +00:00
Graham Fagg
8667648a1b Simple fix (for ticket 363). We push segment size to type size. In other algorithms we switch of segementing altogether. But really the DDT can probably handle partial types so we could really keep the segsize constant (for all but reduce ops) and treat it just as byte arrays..
todos: macroize it as we do it 10 different ways, add mca params to control handling (push up size, no change, switch off segmenting)

This commit was SVN r11619.
2006-09-12 00:01:27 +00:00
George Bosilca
3b39df8ae1 More protection around what we really want to get exported.
This commit was SVN r11437.
2006-08-27 04:49:02 +00:00
George Bosilca
3f0a7cad9e The last patch for Windows support. Mostly casting and conversion to C++ friendly headers.
This commit was SVN r11400.
2006-08-24 16:38:08 +00:00
George Bosilca
6afa4c6c64 Windows friendly version. We have to split the OMPI_DECLSPEC in at least 3
different macros, one for each project. Therefore, now we have OPAL_DECLSPEC,
ORTE_DECLSPEC and OMPI_DECLSPEC. Please use them based on the sub-project.

This commit was SVN r11270.
2006-08-20 15:54:04 +00:00
Graham Fagg
f10c21b746 corrected mca param description and algorithm count
(now to find out why I have disallowed direct calling fo the bm tree)

This commit was SVN r10603.
2006-06-30 23:22:49 +00:00
Graham Fagg
f64cbbe8f2 ops. some decisions used extent rather than size for decision making
yes this means it WAS possible for two nodes to choice two different algorithms
(discovered by Doug Gregor and figured out by George)
Also changed some names like size to comsize so we know which sizes we are using where
This should be updated in al versions

This commit was SVN r10601.
2006-06-30 21:49:04 +00:00
Graham Fagg
c31a5ad4b3 A few small changes that just expanded in the name of neatness...
(1) As pointed out by Torsten after Jeff comment that there are 15 collectives yesterday.. nope.. I have 16 but
    miss counted them in my ifdefs (I had two #11s). Replaces with enum...
(2) Added a readonly MCA param for how many backend algorithms are available per collective (used by benchmarker/STS)
    This allowed me to remove the tuned query internal functions and replace them with ompi_coll_tuned_forced_max_algorithms[COLL].
(3) I was reading the user forced MCA params for the collectives on each comm create (module init) but I then put the 
    values into a global set of variables (like ompi_coll_tuned_reduce_forced_algorithm).

    To fix this and make the code neater:
    (a) The component looks up the MCA param indices on Open if dynamic_rules is set via the
                        ompi_coll_tuned_COLLECTIVE_intra_check_forced_init () call.
    (b) Got rid of the ompi_coll_ompi_coll_tuned_COLLECTIVE_forced_algorithm/segmentsize/etc globals with a struct that
            is now cached on the module data hung off the communicator. i.e. done right.
    (c) On module init if dynamic rules enabled we call a general getvalues routine (in coll_tuned_forced.c) to get the
            CURRENT values using the MCA param indices and then put them on the modules data segment.
        A shorter version of getvalues exists for barrier which only needs the algorithm choice

This commit was SVN r9663.
2006-04-19 23:42:06 +00:00
Graham Fagg
95b060c741 output the right name and stop confusing george
This commit was SVN r9215.
2006-03-08 00:40:14 +00:00
George Bosilca
39252b764f Correctly compute the size of the datatype.
This commit was SVN r9127.
2006-02-23 04:30:52 +00:00
George Bosilca
805c45de29 Don't let a division by zero happens ...
This commit was SVN r9109.
2006-02-22 06:34:05 +00:00
Brian Barrett
566a050c23 Next step in the project split, mainly source code re-arranging
- move files out of toplevel include/ and etc/, moving it into the
    sub-projects
  - rather than including config headers with <project>/include, 
    have them as <project>
  - require all headers to be included with a project prefix, with
    the exception of the config headers ({opal,orte,ompi}_config.h
    mpi.h, and mpif.h)

This commit was SVN r8985.
2006-02-12 01:33:29 +00:00
Graham Fagg
232bb9534a Start moving stuff out of modules that should be in the component.
This commit was SVN r8874.
2006-02-01 20:50:14 +00:00
Graham Fagg
5f2d82347f a couple of changes to make barrier synchronous.. means last communication to any possible peer must
be locally completing. for now using synchronous calls until the new functionality is available. then will change
the code to use the new PML send flags.

This commit was SVN r8867.
2006-01-31 23:21:46 +00:00
Graham Fagg
25375759c3 arrgh. reduce could for very small message sizes and proc counts call a linear function
this was implemented using a chain (tree followed with pipeline) by setting the chain fanout to a factor of size etc but the chain datastructure was fixed in length and if exceeded the topo create returned a null which isn't helpfull in cid next function of comdup...
Anyway two fixes, first we do have a real linear function so changed the decision function and second altered the
topo chain create to force chain fanouts of less than 1 to 1 and fanouts bigger than max to max.
next check in will change chain to dynamically allocd array (reallocable) but we shouldn't ever use a chain fanout for a linear tree anyway. 
(lession must rerun all tests for all data sizes when changing decision functions)

This commit was SVN r8662.
2006-01-08 02:41:09 +00:00
Jeff Squyres
54c4bd3ce2 Update to have public symbols be consistent; use new prefix rule
(apparently we've been doing this in opal and orte, but not in ompi
yet).  All public symbols begin with "ompi_coll_tuned_" (not
mca_coll_tuned_) except the component struct.  Now this component
passes the illegal symbol report with no hits.

This commit was SVN r8589.
2005-12-22 13:49:33 +00:00
Jeff Squyres
2435970cb8 Enable the new "tuned" coll component in an attempt to get wider
testing.  Note that this effectively replaces the "basic" component as
the baseline collective component.  Please report any problems with
this component.

If you run into problems with this component, you can disable it with:

   --mca coll_tuned_priority 0

This commit was SVN r8575.
2005-12-21 12:43:03 +00:00
Graham Fagg
8651658816 minor compile warnings fix
This commit was SVN r8497.
2005-12-14 19:09:46 +00:00
George Bosilca
b7353c707d Remove unprotected header files.
This commit was SVN r8432.
2005-12-10 17:04:46 +00:00
Graham Fagg
141d4ea30a cleaned up ready for changes to move cached data off the MCW module to the
component (where they belong)

This commit was SVN r8407.
2005-12-08 03:14:57 +00:00
George Bosilca
1aa6d27ffe Remove all the compilation warnings I found including unused variables and functions.
This commit was SVN r8226.
2005-11-22 03:42:15 +00:00
George Bosilca
2749870f2c I'm in for the tuned collective module :)
This commit was SVN r8146.
2005-11-13 23:04:14 +00:00
George Bosilca
d8d13f879f When --disable-debug is specified we have to explicitly include the optl/util/output.h header.
This commit was SVN r8133.
2005-11-12 04:03:19 +00:00
Graham Fagg
877f7bbe6a File based dynamic up and tested...
Lots of misc fixes: printfs->opal_output, handles fanin/out correctly for forced ops
unused vars, correct calculations on meaning of 'msgsize' for decision functions
(varies depending on algorithm), etc

This commit was SVN r8113.
2005-11-11 04:49:29 +00:00
George Bosilca
3507d5e9cd opal/util/output.h is required for optimized builds.
This commit was SVN r8076.
2005-11-10 01:19:27 +00:00
Graham Fagg
6b99301893 extra verbose in debug mode to help occ
This commit was SVN r8061.
2005-11-09 21:01:35 +00:00
Graham Fagg
bcf8744bf6 valgrind saved me from a nasty order of eval error... i.e. derefing slected_data before setting it.
Anyway fixed and no memory leaks in coll tuned so far.

This commit was SVN r8037.
2005-11-08 04:52:30 +00:00
Graham Fagg
5b3ba944a8 Enabled, and running...
todos. turn the debug messages  into ompi ignorables and inot do some ops in ompi_bug mode

This commit was SVN r8036.
2005-11-08 04:43:17 +00:00
Graham Fagg
833b558046 Full configuration file based control of tuned collectives.
(verbose on bad config file and even cleans up after itself enought to make valgrind happy).

This commit was SVN r8035.
2005-11-08 03:36:38 +00:00
Graham Fagg
39207db7cd removed the n-dimmension rule base..
replacing it was simpler code for V1

This commit was SVN r8033.
2005-11-08 03:03:51 +00:00
Graham Fagg
dcd3450e06 simplified the building of different rule sets
(also corrected some prototypes missing 'struct')

This commit was SVN r8003.
2005-11-06 22:05:50 +00:00
Jeff Squyres
42ec26e640 Update the copyright notices for IU and UTK.
This commit was SVN r7999.
2005-11-05 19:57:48 +00:00
Graham Fagg
9547a635a9 snapshot while switching systems
but, dynamic rules from a user defined config file is almost there now

This commit was SVN r7943.
2005-11-01 00:19:05 +00:00
Graham Fagg
fe03e068f2 allow forced algorithms (where the user or *test* suite knows better) to
go through the dynamic decision rule interface.
(forced algorithms are set with MCA params)
fixed some silly verbose output with wrong func name in it etc
updates to fixed dec rules.

This commit was SVN r7940.
2005-10-31 20:45:50 +00:00
Graham Fagg
5bb0d4a053 enable allreduce to be selected
This commit was SVN r7888.
2005-10-26 23:55:37 +00:00
Graham Fagg
2587d7ade9 added some more linear functions.
minor corrections on naming and debug info

This commit was SVN r7887.
2005-10-26 23:51:56 +00:00
Graham Fagg
c3e1dc410d Started to add basic linear functions
Also started to add the allreduce algorithms as I test them
(i.e. if it goes in its after testing from now on)

This commit was SVN r7886.
2005-10-26 23:11:32 +00:00
Graham Fagg
382f05c7ad Infastructure changes.
started to add static (fixed if) statement based decision rules based on gigE numbers
added mca params so that a user can force a certain algorithm/segment/topo on a per collective basis
(this is not in the fixed call path but only in the dynamic (at com create) call path).
(these params can be used by test suites such as OCC to choice which algorithm they are using).

This commit was SVN r7854.
2005-10-25 03:55:58 +00:00
Graham Fagg
d8e32464cb ops. setting/reading mca option from the right varible would help.
This commit was SVN r7850.
2005-10-24 21:33:48 +00:00
Graham Fagg
636b42afff handle non existant recv buf in reduce for non root processes
(basic allreduce does this for mpi_in_place case)

This commit was SVN r7759.
2005-10-14 00:00:37 +00:00
Graham Fagg
61b8218d76 MPI_IN_PLACE fix for reduce.
(actually a work around for an optimisation in the reduce for not saving ops on the first recv of each segment)
Minor change in topo.

This commit was SVN r7758.
2005-10-13 23:38:21 +00:00
Graham Fagg
607bdf51b6 Last Cleanup BEFORE adding last two methods and final cross over points.
- new mca param calls
- move printfs to OPAL_OUTPUT

This commit was SVN r7692.
2005-10-11 18:51:03 +00:00
Jeff Squyres
15d0a95202 - Remove extra whitespace from Makefile.am's from when we removed
Makefile.options
- Sample in each of the three projects of how to link againt the
  relevant libraries so that when components are loaded into a parent
  process' space, we don't rely on the libopal/liborte/libmpi symbols
  being in the parent's public symbol namespace -- instead,
  dynamically link to the relevant libraries, allowing the dynamic
  linker to pull those libraries in at run-time, if needed

This commit was SVN r7397.
2005-09-15 20:56:18 +00:00
Graham Fagg
0f75381e56 Added various barrier routines: recursive doubling, bruck, double ring, 2proc etc
all pass tests

This commit was SVN r7355.
2005-09-13 20:58:42 +00:00
Graham Fagg
9053790973 fixed bruck alltoall bug. now passes ibm tests
This commit was SVN r7346.
2005-09-13 18:35:45 +00:00
Graham Fagg
f354140587 Multiple changes
-added some alltoall calls (pairwise checked ok, bruck testing)
-changes in use of data hung of communicator
-making sendrecv call a true inline function
-more use ompi_ddt routines

This commit was SVN r7337.
2005-09-13 04:28:18 +00:00
Graham Fagg
00842ad87c Fixed reduce bugs
(added some temp debug routines)

This commit was SVN r7283.
2005-09-09 23:05:17 +00:00
Graham Fagg
9596d6e2f3 use correct names or cannot call functions...
use correct ddt calls (and params ext vs lb etc)

This commit was SVN r7243.
2005-09-09 02:39:57 +00:00
Graham Fagg
b97305bbcc a few more functions (testing)
This commit was SVN r7190.
2005-09-06 05:21:57 +00:00
Graham Fagg
059d7f26f8 re-enable proper freeing of allocated structures now that I know these
are not causing an error.

This commit was SVN r7170.
2005-09-03 06:19:17 +00:00
Graham Fagg
35447d4c32 underfilling of a only segment and the doing a PML send and esp a recv on it can
cause flipping hard to find memory errors...

This commit was SVN r7169.
2005-09-03 06:14:03 +00:00
Graham Fagg
36eddb6609 checkpoint
This commit was SVN r7168.
2005-09-03 01:41:13 +00:00
Graham Fagg
25787c4ca7 added segmented binary
renamed split so we know its a split binary tree
(tested, chain/bin/splitbin all pass simple and ibm tests)

This commit was SVN r7138.
2005-09-02 02:19:59 +00:00
Graham Fagg
a7a20e1a57 corrected first seg recv for intermediate nodes (or else we lose a request)
This commit was SVN r7137.
2005-09-02 00:55:47 +00:00
Graham Fagg
d78e9d0e9e get/handle extent correctly
last rank in comm is not always last in tree structure for nonzero roots

This commit was SVN r7135.
2005-09-01 22:56:47 +00:00
Graham Fagg
f4de8776a6 ops
This commit was SVN r7133.
2005-09-01 17:12:28 +00:00
Graham Fagg
4ac2445c61 Various clean ups
Changed component so choice of decision functions controlled by mca params
(for now fixed decision functions (if statements) default)
started fixes for the various bcasts

This commit was SVN r7117.
2005-08-31 23:38:09 +00:00
Graham Fagg
1caec16018 snapshot as I move machines
do not use

This commit was SVN r7103.
2005-08-31 01:43:48 +00:00