- 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.
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.
is that if one add "pml=" to the configuration file, really bad things
happen. All PMLs will get initialize, and each of them will initialize
all BTLs. This patch force the mca_pml_base_pml to get initialized in
all cases before we go out of the mca_pml_base_open function.
This commit was SVN r12527.
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.
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.
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.
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.
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.
This commit essentially caches the invoking comm/win/file on the
ompi_request_t. This, paired with the req_type field, allows us to
retrieve the invoking MPI object and invoke the proper errhandler.
The patch is missing most updates for the MPI-2 one-sided stuff (i.e.,
the patch mainly fixes comms and files); I didn't really understand
that code and didn't want to hazard trying to figure it out when Brian
can probably do it much more quickly.
So #250 will still stay open, pending MPI-2 one-sided updates for this
stuff.
This commit was SVN r12339.
The following Trac tickets were found above:
Ticket 250 --> https://svn.open-mpi.org/trac/ompi/ticket/250
* Create a new request type: NOOP (described below)
* For all MPI_*_INIT functions, OBJ_NEW an ompi_request_t and set its
type to NOOP
* Ensure that the NOOP requests are OBJ_RELEASE'd when they are done
* MPI_START looks at the request type; if NOOP, just return success. If
not, call the PML start() function
* MPI_STARTALL always pass the entire array of requests back to the PML
(see next point)
* Make the PMLs only process PML requests (i.e., ignore/skip anything
that isn't of type PML -- such as the NOOP requests)
* Add a little more param error checking in STARTALL
This commit was SVN r12338.
The following Trac tickets were found above:
Ticket 529 --> https://svn.open-mpi.org/trac/ompi/ticket/529
allocation logic is completely done outside the data-type engine (in the PML) there is
no need for any special case inside the data-type engine. There is less arguments for
the ompi_convertor_pack and ompi_convertor_unpack as well (the last field free_after is
not required anymore as there is no memory allocated in the engine itself). This change
affect all components using datatypes. I test most of them, but it might happens that I
miss some ... If it's the case please let me know (don't shoot the pianist!!).
This commit was SVN r12331.
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.
A segfault would occur in mca_pml_ob1_recv_request_progress() when trying to prepare the convertor for unpacking, because the request's req_proc field was NULL.
Turns out that we weren't setting the req_proc field in the MCA_PML_OB1_CHECK_SPECIFIC_AND_WILD_RECEIVES_FOR_MATCH macro. Instead of just setting it there I removed the other place req_proc was being set correctly, and instead took care of all the cases at once in mca_pml_ob1_recv_frag_match().
This commit was SVN r12323.
parameter. For optimisation purpose only this BTL is used to send packet
through instead of trying to send packets through all BTLs. But actually the
code was wrong. It simply used provided bml_btl and it may represent different
endpoint from packet's destination. The fixed code checks if packet's
destination is reachable through the BTL, finds appropriate bml_btl and only
then tries to send it through correct bml_btl.
This commit was SVN r12319.
is done to assure alignment so strictly aligned CPUs (like SPARC) do not
sigbus. This also may benefit other platforms too.
This commit fixes trac:494.
This commit was SVN r12312.
The following Trac tickets were found above:
Ticket 494 --> https://svn.open-mpi.org/trac/ompi/ticket/494
mentioned in the comment the completion/callback of the triggered
send operation can happen before the call returns. If this happens and
if the pipeline depth is 0 before we triggered the send operation and
this is the last send operation of the request then the completion detection
code will decrement the pipeline depth and check it for equality to 0.
Because (0-1) != 0 the pml completion function for this request will
*not* be called.
This part 2 of the fix for ticket #246.
This commit was SVN r12292.
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.
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.
The UD BTL isn't gone - the latest version is in my afriedle-ud branch. This version on the trunk was very old, ompi_ignore'd, lacked performance, and probably contained bugs. The maintained version on my branch is working solid, and will eventually come back, but not for v1.2.
This commit was SVN r12144.
where a window was in both the passive and active side of a lock sequence.
Refs trac:488
This commit was SVN r12112.
The following Trac tickets were found above:
Ticket 488 --> https://svn.open-mpi.org/trac/ompi/ticket/488
constrained:
* Make sure we always have a number of eager fragments available
that scales with the number of processes communicating with
a given proc over shared memory
* Use FREE_LIST_GET instead of FREE_LIST_WAIT to return an
error to the PML when resource exhaustion occurs
* Don't dereference the frag during alloc unless we're sure
it's not NULL
Reviewed by: Galen
Refs trac:413
This commit was SVN r12053.
The following Trac tickets were found above:
Ticket 413 --> https://svn.open-mpi.org/trac/ompi/ticket/413
First, move the OPAL_THREAD_LOCK out to the same level as its corresponding UNLOCK. It was possible to hit the UNLOCK without ever acquiring the lock.
Since the OPAL_THREAD_ADD64() is now protected by this lock, we can just do the decrement non-atomically.
This commit was SVN r11958.
Don't try to acquire ompi_request_lock here, which in all cases is already held. Avoids deadlock that occurs even when threads are enabled and we're running a THREAD_SINGLE app.
Reviewed by Galen.
This commit was SVN r11957.
The following Trac tickets were found above:
Ticket 183 --> https://svn.open-mpi.org/trac/ompi/ticket/183
I do something else" rule screws me up again. If we're in a FENCE, but
not in ACCESS | EXPOSE, put us in ACCESS|EXPOSE, as we are now known we
now in a real Fence epoch. Yay silly MPI standards
Refs trac:441
This commit was SVN r11865.
The following Trac tickets were found above:
Ticket 441 --> https://svn.open-mpi.org/trac/ompi/ticket/441
on the number of known local procs, with a high and low watermark. Right
now, we default to a low watermark point of 64MB, a per-proc scaling
factor of 32MB, and a high watermark point of 512MB
Refs trac:212
This commit was SVN r11824.
The following Trac tickets were found above:
Ticket 212 --> https://svn.open-mpi.org/trac/ompi/ticket/212
GIDs (there can be more than one) and not GIDs of the HCA on the network. Entry
zero always have to be initialized so we use it, and warn user if there is more
then one port active and default subnet is configured on at least one of them.
This commit was SVN r11815.
the component is configured successfully. Otherwise, we can end up
trying to run make in the romio directory without any Makefiles. This
really only happens on the targets that recurse into DIST_SUBDIRS - ie
dist, maintainer-clean, and distclean
refs trac:411
This commit was SVN r11807.
The following Trac tickets were found above:
Ticket 411 --> https://svn.open-mpi.org/trac/ompi/ticket/411
tell if the remote proc should be in an exposure epoch or not.
Refs trac:325
This commit was SVN r11746.
The following Trac tickets were found above:
Ticket 325 --> https://svn.open-mpi.org/trac/ompi/ticket/325
epoch's control data could overwrite the previous epoch's data because
we were reusing data structures between PW and SC. Instead, we now
have explicit post_msg and complete_msg counters for completion.
refs trac:354
* Only register the rdma osc callback once, as it turns out that some
btls (MX) do somethng more than update a table during the register
call, and each register call sucks up valuable fragments...
This commit was SVN r11745.
The following Trac tickets were found above:
Ticket 354 --> https://svn.open-mpi.org/trac/ompi/ticket/354
long ago) supposed to be used as a cache for accessing the PML procs. But in
all of the PMLs the PML proc contain only one field i.e. a pointer to the ompi_proc.
This pointer can be accessed using the c_remote_group easily. Therefore, there is no
meaning of keeping the PML procs around. Slim fast commit ...
This commit was SVN r11730.
if we want to be able to reuse the request. If not, the request will never be freed
even if the user call MPI_Request_free.
This commit was SVN r11717.
This provides support for the Infinipath interconnect using the PSM API.
Of note:
This version has a "hackaround" we always return 1 or greater from
the MTL PSM progress function, this should be examined further.
This commit was SVN r11655.
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
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.
* Print a warning error message if a target is not in an exposure epoch
and an update is received. This results in the app continuing with
that call having never happened, rather than evil hangs.
refs trac:325
This commit was SVN r11514.
The following Trac tickets were found above:
Ticket 325 --> https://svn.open-mpi.org/trac/ompi/ticket/325
bunch of code changed indenting level and some code got moved out of
one function and made into its own subroutine.
- Gleb pointed out that I wasn't taking into account values from the
default section of the INI file (and not finding values in the INI
file is not an error).
- I incorrectly thought that 0x5ad was Mellanox's vendor ID. Turns
out that 0x5ad is Cisco's ID, while 0x2c9 is Mellanox.
Specifically, Cisco burns its own firmware into the HCA which
replaces the vendor ID, although the part ID stays the same. So
it's Mellanox hardware with Cisco firmware. And apparently several
of us do that. :-) So I expanded the concept of the vendor_id in
the INI file to allow for lists of vendor IDs.
- Along with that, I updated the default INI file to list all the IB
vendors (that I am aware of -- certainly open to putting more data
in there from other vendors) who overwrite Mellanox's vendor_id with
their own for the part numbers that we have on file.
This commit was SVN r11506.
can be in both a Post and Start state. Also, the asserts were only
correct assuming that we were never in the post and start state at the
same time, which was obviously silly.
refs trac:303
This commit was SVN r11428.
The following Trac tickets were found above:
Ticket 303 --> https://svn.open-mpi.org/trac/ompi/ticket/303
I know it does not make much sense but one can play around with the
performance. Numbers are available at http://www.unixer.de/research/nbcoll/perf/.
This is the first step towards collv2. Next step includes the addition
of non-blocking functions to the MPI-Layer and the collv1 interface.
It implements all MPI-1 collective algorithms in a non-blocking manner.
However, the collv1 interface does not allow non-blocking collectives so
that all collectives are used blocking by the ompi-glue layer.
I wanted to add LibNBC as a separate subdirectory, but I could not
convince the buildsystem (and had not the time). So the component looks
pretty messy. It would be great if somebody could explain me how to move
all nbc*{c,h}, and {hb,dict}*{c,h} to a seperate subdirectory.
It's .ompi_ignored because I did not test it exhaustively yet.
This commit was SVN r11401.
sbrk and the use of mmap(). So rather than checking just for mallopt(),
we should also be checking for those defines when determining if we can
disable giving memory back to the OS or not.
This commit was SVN r11279.
allocated directly by the OS in the paging file (the HUGE file
that cannot be defragmented with any tools). Unlike UNIX, they
do not have physical existence as files.
This commit was SVN r11273.
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.
In order to provide backwards compatability the framework versions are bumped
and the handler registeration function is at the end of the btl struct.
Testing done on sm, openib, and gm..
This commit was SVN r11256.
through the c_pml_procs, as it might be an intercommunicator and therefore
c_my_rank might not be a valid index.
Fixes trac:266.
This commit was SVN r11238.
The following Trac tickets were found above:
Ticket 266 --> https://svn.open-mpi.org/trac/ompi/ticket/266
so use the req_buff field for keeping track of the bsend buffer and the
req_addr field for the user buffer, the way the comments suggested we
were doing it
This commit was SVN r11233.
the upperlayer assynchronously although there are some issues with this.. such
as there are multiple consumers of the btl's.. who get's the
This commit was SVN r11232.
