Aurelien:
What is the cause of this performance penalty?
Jody
On Fri, Apr 11, 2008 at 1:44 AM, Aurélien Bouteiller
wrote:
> Open MPI can manage heterogeneous system. Though you prefer to avoid
> this because it has a performance penalty. I suggest you compile on
> the 32bit machine and use t
In Open MPI, we detect when the source and receiver side share the
same architecture. Then, the datatype engine avoids type conversion
(let say XDR, ou htonl) and sends raw binary data on the link. If the
architecture mismatch, the datatype engine is forced to use some
conversion which will
Hello,
I am a new user of MPI and I would like to have your opinion on the
following issue.
With 1 processor I keep the data in
std::vector > X
Then I call a function that takes various arguments, among which X[ i ] and
X[ j ], i.e. two sub-vectors of X.
With MPI I would have the same function
Hi Alessandro,
As far as I understood , you want to send a row of your two dimensional
array using MPI sending functions.
This is of course possible in a simple way. You can send/receive data
with any type and any size using MPI
but you should tell MPI the type your data using MPI_Datatype stru
On Apr 4, 2008, at 2:47 PM, Matt Hughes wrote:
I was able to eliminate the hang I was seeing with 1.2.5 during the
gather operation by using these btl parameters (found at
http://svn.open-mpi.org/trac/ompi/browser/trunk/ompi/mca/btl/openib/btl-openib-benchmark)
:
btl_openib_max_btls=20
btl_ope
On Apr 6, 2008, at 5:58 PM, Teng Lin wrote:
In order to run a 32-bit program on a 64-bit cluster, one has to build
32-bit OpenMPI. Following some instructions on this mailing list, I
successfully built OpenMPI 1.2.4 on 64-bit OS.
It's not clear from your text -- did you build Open MPI as 32 or
Also, it's been a long time since I've done anything serious in C++,
but are you guaranteed that &foo[x] for a vector<> will yield
contiguous memory?
You might be better off using the Boost C++ MPI bindings; they provide
all kinds of nifty things for classes, etc.
On Apr 7, 2008, at 7:47
