The reason so many different distributions are described is because
what is optimal depends so much on your own case.
Even if one disregards CYCLIC axes, there are still all those BLOCK
choices you mention. It isn't just a matter of choosing which axes
will be * since * is just a special case
Two comments interwoven below...
Cristobal Navarro wrote:
i was not aware that openMPI internally uses shared memory in case two
proceses reside on the same node,
which is perfect.
The ways OMP uses shared memory and Open MPI (or most other MPI
implementations) uses shared memory are very dif
Yves Caniou wrote:
Le Wednesday 28 July 2010 15:05:28, vous avez écrit :
I am confused. I thought all you wanted to do is report out the binding of
the process - yes? Are you trying to set the affinity bindings yourself?
If the latter, then your script doesn't do anything that mpi
I can't give you a complete answer, but I think this is less an MPI
question and more of a Fortran question. The question is if you have a
Fortran derived type, one of whose components is a POINTER, what does
the data structure look like in linear memory? I could imagine the
answer is impleme
christophe petit wrote:
Thanks for your answers,
the execution of this parallel program works fine at my work, but we
used MPICH2. I thought this will run with OPEN-MPI too.
In your input deck, how big are x_domains and y_domains -- that is,
iconf(3) and iconf(4)? Do they have to be chan
Mark Potts wrote:
Hi,
I'd opt for the fact that tv0 is given value only on rank 0 and
tv1 is
only given value on rank 1. Kind of hard to get a diff betwn the two
on either rank with that setup. You need to determine the tv0 and
tv1
on both ranks.
I don't understand this.
Tad Lake wrote:
I run it :
mpirun -np 2 --host node2 ./a.out
But the result of time is not stable with difference of 100 times. For example,
the max value of time can be 3000, meanwhile the min is 100.
Again, know what results to expect. Is 3000 a reasonable time and 100
too fast? Or, is
David Akin wrote:
All,
I'm trying to get the OpenMP portion of the code below to run
multicore on a couple of 8 core nodes.
I was gone last week and am trying to catch up on e-mail. This thread
was a little intriguing.
I agree with Ralph and Terry:
*) OMPI should not be binding by defaul
Eugene Loh wrote:
Yves Caniou wrote:
Le Wednesday 28 July 2010 15:05:28, vous avez écrit :
I am confused. I thought all you wanted to do is report out the binding of
the process - yes? Are you trying to set the affinity bindings yourself?
If the latter, then your
Personally, I've been having trouble following the explanations of the
problem. Perhaps it'd be helpful if you gave us an example of how to
reproduce the problem. E.g., short sample code and how you run the
example to produce the problem. The shorter the example, the greater
the odds of reso
The way MPI processes are being assigned to hardware threads is perhaps
neither controlled nor optimal. On the HT nodes, two processes may end
up sharing the same core, with poorer performance.
Try submitting your job like this
% cat myrankfile1
rank 0=os223 slot=0
rank 1=os221 slot=0
rank
sun...@chem.iitb.ac.in wrote:
sun...@chem.iitb.ac.in wrote:
Dear Open-mpi users,
I installed openmpi-1.4.1 in my user area and then set the path for
openmpi in the .bashrc file as follow. However, am still getting
following
error message whenever am starting the paralle
Terry Dontje wrote:
Jeff Squyres wrote:
You should be able to run "./configure --help" and see a lengthy help message that includes all the command line options to configure.
Is that what you're looking for?
No, he wants to know what configure options were used with some
binarie
Gus Correa wrote:
More often than not some components lag behind (regardless of how
much you tune the number of processors assigned to each component),
slowing down the whole scheme.
The coupler must sit and wait for that late component,
the other components must sit and wait for the coupler,
an
Alex A. Granovsky wrote:
Isn't in
evident from the theory of random processes and probability theory that in the limit of infinitely
large cluster and parallel process, the probability of deadlocks with current implementation is unfortunately
quite a finite quantity and in
Richard Treumann wrote:
Hi Ashley
I understand the problem with
descriptor
flooding can be serious in an application with unidirectional data
dependancy.
Perhaps we have a different perception of how common that is.
Ashley speculated it was a "significant minority." I don't know
Georges Markomanolis wrote:
I have some questions about the duration of the communication with
MPI_Send and MPI_Recv. I am using either SkaMPI either my
implementation to measure the pingpong (MPI_Send and MPI_Recv) time
between two nodes for 1 byte and more. The timing of the pingpong is
106
Georges Markomanolis wrote:
Dear all,
Hi again, after using MPI_Ssend seems to be what I was looking for but
I would like to know more about MPI_Send.
For example sending 1 byte with MPI_Send it takes 8.69 microsec but
with MPI_Ssend it takes 152.9 microsec. I understand the difference
but
It seems to me there are two extremes.
One is that you replicate the data for each process. This has the
disadvantage of consuming lots of memory "unnecessarily."
Another extreme is that shared data is distributed over all processes.
This has the disadvantage of making at least some of the
Storm Zhang wrote:
Here is what I meant: the results of 500 procs in fact shows it with
272-304(<500) real cores, the program's running time is good, which is
almost five times 100 procs' time. So it can be handled very well.
Therefore I guess OpenMPI or Rocks OS does make use of hyperthread
David Ronis wrote:
Ralph, thanks for the reply.
If I build with enable-heterogeneous and then decide to run on a
homogeneous set of nodes, does the additional "overhead" go away or
become completely negligible; i.e., if no conversion is necessary.
I'm no expert, but I think the overhead d
Ed Peddycoart wrote:
After searching some more and reading some FAQs on the
opmi website, I see sugestions on how to make a remote app use the
local display to render, but that isn't what I need... Let me revise or
clarify my question:
I have an app which will run on 5 mach
Ed Peddycoart wrote:
I need to do some performance tests on my mpi app. I simply want
to determine how long it takes for my sends from one process to be
received by another process.
Here is the code I used as my example for non-blocking send/receive...
if( myrank == 0 ) {
/*
Brian Budge wrote:
Hi all -
I just ran a small test to find out the overhead of an MPI_Recv call
when no communication is occurring. It seems quite high. I noticed
during my google excursions that openmpi does busy waiting. I also
noticed that the option to -mca mpi_yield_when_idle seems no
Jeff Squyres wrote:
Ah. The original code snipit you sent was:
MPI::COMM_WORLD.Barrier();
if(rank == master) t1 = clock();
"code A";
MPI::COMM_WORLD.Barrier();
if(rank == master) t2 = clock();
"code B";
Remember that the time that individual processes exit barrier is not guaranteed
to be uni
n it in more
details for the second sentence. Thanks a lot.
On Thu, Oct 21, 2010 at 9:58 AM, Eugene Loh <eugene@oracle.com>
wrote:
Jeff Squyres wrote:
MPI::COMM_WORLD.Barrier();
if(rank == master) t1 = clock();
"code A";
MPI::COMM_WORLD.Barrier();
if(rank == m
christophe petit wrote:
i'm studying the parallelized version of a solving 2D heat equation
code in order to understand cartesian topology and the famous
"MPI_CART_SHIFT".
Here's my problem at this part of the code :
---
jody wrote:
Where is the option 'default-hostfile' described?
Try "mpirun --help". Not everything makes it to the man page. Heck,
not everything is documented!
It does not appear in mpirun's man page (for v. 1.4.2)
and i couldn't find anything like that with googling.
On Wed, Oct 27, 2
christophe petit wrote:
i
am still trying to understand the parallelized version of the heat
equation 2D solving that we saw at school.
I am confused between the shift of the values near to the bounds done
by the "updateBound" routine and the main loop (at line 161 in main
code) which c
Debugging is not a straightforward task. Even posting the code doesn't
necessarily help (since no one may be motivated to help or they can't
reproduce the problem or...). You'll just have to try different things
and see what works for you. Another option is to trace the MPI calls.
If a proc
Try lowering the eager threshold more gradually... e.g., 4K, 2K, 1K,
512, etc. -- and watch what happens. I think you will see what you
expect, except once you get too small then the value is ignored
entirely. So, the setting just won't work at the extreme value (0) you
want.
Maybe the thin
George Markomanolis wrote:
Dear Eugene,
Thanks a lot for the answer you were right for the eager mode.
I have one more question. I am looking for an official tool to measure
the ping time, just sending a message of 1 byte or more and measure
the duration of the MPI_Send command on the rank 0
Gilbert Grosdidier wrote:
I found this parameter mpool_sm_max_size in this post:
http://www.open-mpi.org/community/lists/devel/2008/11/4883.php
But I was unable to spot it back into the 'ompi_info -all'
output for v 1.4.3.
Is it still existing ?
No.
If not, which
Ioannis Papadopoulos wrote:
Has anyone observed similar behaviour? Is it something that I'll have
to deal with it in my code or does it indeed qualify as an issue to be
looked into?
I would say this is NOT an issue that merits much attention. There are
too many potential performance anomali
Mathieu Gontier wrote:
Dear OpenMPI users
I am dealing with an arithmetic problem. In fact, I have two variants
of my code: one in single precision, one in double precision. When I
compare the two executable built with MPICH, one can observed an
expected difference of performance: 115.7-se
Mathieu Gontier wrote:
Nevertheless, one can observed some differences between MPICH and
OpenMPI from 25% to 100% depending on the options we are using into
our software. Tests are lead on a single SGI node on 6 or 12
processes, and thus, I am focused on the sm option.
Is it possible to narr
Ralph Castain wrote:
I know we have said this many times - OMPI made a design decision to poll hard
while waiting for messages to arrive to minimize latency.
If you want to decrease cpu usage, you can use the yield_when_idle option (it
will cost you some latency, though) - see ompi_info --par
Terry Dontje wrote:
On 12/10/2010 09:19 AM, Richard Treumann wrote:
It seems to me the MPI_Get_processor_name
description is too ambiguous to make this 100% portable. I assume most
MPI implementations simply use the hostname so all processes on the
same host will return the same string.
David Mathog wrote:
Also, in my limited testing --host and -hostfile seem to be mutually
exclusive.
No. You can use both together. Indeed, the mpirun man page even has
examples of this (though personally, I don't see having a use for
this). I think the idea was you might use a hostfile to
Alaukik Aggarwal wrote:
Hi,
I am using Open MPI 1.4.3.
I have to perform a receive operation from processes that are sending
data. It might happen that some of the processes don't send data
(might have completed in-fact).
So, how do I perform check on which processes to receive data from and
charge of defining MPI calls.
On Sat, Dec 11, 2010 at 10:28 AM, Eugene Loh wrote:
Alaukik Aggarwal wrote:
I am using Open MPI 1.4.3.
I have to perform a receive operation from processes that are sending
data. It might happen that some of the processes don't send data
(might ha
David Mathog wrote:
Is there a rule of thumb for when it is best to contact N workers with
MPI_Bcast vs. when it is best to use a loop which cycles N times and
moves the same information with MPI_Send to one worker at a time?
The rule of thumb is to use a collective whenever you can. The
ra
David Mathog wrote:
Is there a tool in openmpi that will reveal how much "spin time" the
processes are using?
I don't know what sort of answer is helpful for you, but I'll describe
one option.
With Oracle Message Passing Toolkit (formerly Sun ClusterTools, anyhow,
an OMPI distribution from
David Mathog wrote:
For the receive I do not see how to use a collective. Each worker sends
back a data structure, and the structures are of of varying size. This
is almost always the case in Bioinformatics, where what is usually
coming back from each worker is a count M of the number of signi
Jeff Squyres wrote:
On Dec 16, 2010, at 5:14 AM, Mathieu Gontier wrote:
We have lead some tests and the option btl_sm_eager_limit has a positive consequence on the performance. Eugene, thank you for your links.
Good!
Just be aware of the tradeoff you're making: space for
Can you isolate a bit more where the time is being spent? The
performance effect you're describing appears to be drastic. Have you
profiled the code? Some choices of tools can be found in the FAQ
http://www.open-mpi.org/faq/?category=perftools The results may be
"uninteresting" (all time sp
I'm curious if that resolved the issue.
David Singleton wrote:
http://www.open-mpi.org/faq/?category=running#oversubscribing
On 12/03/2010 06:25 AM, Price, Brian M (N-KCI) wrote:
Additional testing seems to show that the problem is related to
barriers and how often they poll to determine whe
ture the performance problem you're looking
for?
Le
22/12/2010 18:50, Eugene Loh a écrit :
Can
you isolate a bit more where the time is being spent? The performance
effect you're describing appears to be drastic. Have you profiled the
code? Some choices of tools can be found in the
David Mathog wrote:
Ralph Castain wrote:
Afraid not - though you could alias your program name to be "nice --10 prog"
Is there an OMPI wish list? If so, can we please add to it "a method
to tell mpirun what nice values to use when it starts programs on
nodes"? Minimall
Gilbert Grosdidier wrote:
Any other suggestion ?
Can any more information be extracted from profiling? Here is where I
think things left off:
Eugene Loh wrote:
Gilbert Grosdidier wrote:
#
[time] [calls] <%mpi> <%wall>
#
David Zhang wrote:
Blocking send/recv, as the name suggest, stop processing
your master and slave code until the data is received on the slave side.
Just to clarify...
If you use point-to-point send and receive calls, you can make the
block/nonblock decision independently on the send and rece
Bibrak Qamar wrote:
Hello All,
I am using non-blocking send and receive, and i want to calculate the
time it took for the communication. Is there any method or a way to do
this using openmpi.
You probably have to start by defining what you mean by "the time it
took for the communication".
mmunication.
From: Eugene Loh <eugene@oracle.com>
You probably have to start by defining what you mean by "the time it
took for the communication". Anyhow, the Peruse instrumentation in OMPI
might help.
Again, you should probably start by thinking more precisely about what
tim
jody wrote:
Thanks for your reply.
If i try your suggestion, every process fails with the following message:
*** The MPI_Init() function was called before MPI_INIT was invoked.
That's a funny error message. If you search the OMPI users mail list
archives, this message shows up, but I didn
Correa wrote:
However, OpenMPI may give this info, with non-MPI (hence non-portable)
functions, I'd guess.
From: Eugene Loh
Anyhow, the Peruse instrumentation in OMPI might help.
ll the tools.
Bibrak Qamar wrote:
As asked the reason of such calculation of non
blocking communication, the main reason is that I want to look into the
program as how much it percent time is consumed on communication alone,
computation alone and the intersection of both.
On Thu, Feb 3, 2
Prentice Bisbal wrote:
Jeff Squyres wrote:
Can you put together a small example that
shows the problem...
Jeff,
Thanks for requesting that. As I was looking at the oringinal code to
write a small test program, I found the source of the error. Doesn't it
aways work that way.
No
Eye RCS 51 wrote:
Hi,
In an effort to make a Qt gui using MPI, I have the following:
1. Gui started in master node.
2. In Gui, through a pushbutton, a global variable x is assigned some
value; let say, x=1000;
3. I want this value to be know to all nodes. So I used broadcast in
the functi
Let's say you have multi-threaded MPI processes, you request
MPI_THREAD_MULTIPLE and get MPI_THREAD_MULTIPLE, and you use the
self,sm,tcp BTLs (which have some degree of threading support). Is it
okay to have an [MPI_Isend|MPI_Irecv] on one thread be completed by an
MPI_Wait on another thread?
. That should cover your case. See the man
page for MPI_Init_thread(). My question should not have anything to do
with your case.
On Tue, Mar 8, 2011 at 12:34 PM, Eugene Loh wrote:
Let's say you have multi-threaded MPI processes, you request
MPI_THREAD_MULTIPLE an
Gustavo Correa wrote:
Dear OpenMPI Pros
Is there an MCA parameter that would do the same as the mpiexec switch
'-bind-to-core'?
I.e., something that I could set up not in the mpiexec command line,
but for the whole cluster, or for an user, etc.
In the past I used '-mca mpi mpi_paffinity_alone
Gus Correa wrote:
Ralph Castain wrote:
On Mar 21, 2011, at 9:27 PM, Eugene Loh wrote:
Gustavo Correa wrote:
Dear OpenMPI Pros
Is there an MCA parameter that would do the same as the mpiexec
switch '-bind-to-core'?
I.e., something that I could set up not in the mpiexec command
Michele Marena wrote:
I've launched my app with mpiP both when two processes are
on different node and when two processes are on the same node.
The process 0 is the manager (gathers the results only),
processes 1 and 2 are workers (compute).
This is the case processes 1 and 2 a
Nehemiah Dacres wrote:
also, I'm not sure if I'm reading the results right.
According to the last run, did using the sun compilers (update 1 )
result in higher performance with sunct?
On Wed, Apr 6, 2011 at 11:38 AM, Nehemiah
Dacres
wrote:
this
first test was run as
amosl...@gmail.com wrote:
Hi,
I am embarrassed! I submitted a note to the users on setting
up openmpi-1.4.3 using SUSE-11.3 under Linux and received several
replies. I wanted to transfer them but they disappeared for no
apparent reason. I hope that those that sent me messages wil
CMR 2728 did this. I think the changes are in 1.5.4.
On 4/28/2011 5:00 AM, Jeff Squyres wrote:
It is quite likely that --enable-progress-threads is broken. I think it's even
disabled in 1.4.x; I wonder if we should do the same in 1.5.x...
Depending on what version you use, the option has been renamed
--enable-mpi-thread-multiple.
Anyhow, there is widespread concern whether the support is robust. The
support is known to be limited and the performance poor.
On 5/4/2011 9:14 AM, Mark Dixon wrote:
I've been asked about mixed-mode
On 5/27/2011 4:32 AM, Jeff Squyres wrote:
On May 27, 2011, at 4:30 AM, Robert Horton wrote:
To be clear, if you explicitly list which BTLs to use, OMPI will only
(try to) use exactly those and no others.
It might be worth putting the sm btl in the FAQ:
http://www.open-mpi.org/faq/?category=ope
On 7/12/2011 4:45 PM, Mohan, Ashwin wrote:
I noticed that the exact same code took 50% more time to run on OpenMPI
than Intel.
It would be good to know if that extra time is spent inside MPI calls or
not. There is a discussion of how you might do this here:
http://www.open-mpi.org/faq/?catego
I'm going to echo what you've already heard here: it is impossible for
a process to receive a message that was sent to a different process. A
sender must specify a unique destination. No process other than the
destination process will see that message.
In what you write below, why do you th
On 8/23/2011 12:32 AM, Hoang-Vu Dang wrote:
Dear all,
Where could I find a detailed documentation about algorithms that has
been using in Open MPI ?
For example, I would like to answer following questions: how
MPI_Algather operation is done? what is the complexity in term of the
number of d
On 8/23/2011 1:24 PM, Dick Kachuma wrote:
I have used gprof to profile a program that uses openmpi. The result
shows that the code spends a long time in poll (37% on 8 cores, 50% on
16 and 85% on 32). I was wondering if there is anything I can do to
reduce the time spent in poll.
In serial perfor
On 8/31/2011 11:48 PM, Randolph Pullen wrote:
I recall a discussion some time ago about yield, the Completely F%’d
Scheduler (CFS) and OpenMPI.
My system is currently suffering from massive CPU use while busy
waiting. This gets worse as I try to bump up user concurrency.
Yup.
I am running w
I agree sentimentally with Ghislain. The time spent in a barrier
should conceptually be some wait time, which can be very long (possibly
on the order of milliseconds or even seconds), and the time to execute
the barrier operations, which should essentially be "instantaneous" on
some time scal
On 9/8/2011 7:42 AM, Ghislain Lartigue wrote:
I will check that, but as I said in first email, this strange behaviour happens
only in one place in my code.
Is the strange behavior on the first time, or much later on? (You seem
to imply later on, but I thought I'd ask.)
I agree the behavior i
done=.true.
endif
enddo
The first call to the barrier works perfectly fine, but the second one gives
the strange behavior...
Ghislain.
Le 8 sept. 2011 à 16:53, Eugene Loh a écrit :
On 9/8/2011 7:42 AM, Ghislain Lartigue wrote:
I will check that, but as I sai
On 9/8/2011 11:47 AM, Ghislain Lartigue wrote:
I guess you're perfectly right!
I will try to test it tomorrow by putting a call system("wait(X)) befor the
barrier!
What does "wait(X)" mean?
Anyhow, here is how I see your computation:
A) The first barrier simply synchronizes the processes.
On 9/15/2011 5:51 AM, Ghislain Lartigue wrote:
start_0 = MPI_Wtime()
start_1 = MPI_Wtime()
call foo()
end_1 = MPI_Wtime()
write(*,*) "timer1 = ",end1-start1
start_2 = MPI_Wtime()
call bar()
end_2 = MPI_Wtime()
write(*,*) "timer2 = ",end2-start2
end_0 = MPI_Wtime()
write(*,*) "timer0 = ",end0-s
Should be fine. Once MPI_Send returns, it should be safe to reuse the
buffer. In fact, the return of the call is the only way you have of
checking that the message has left the user's send buffer. The case
you're worried about is probably MPI_Isend, where you have to check
completion with an
On 9/18/2011 9:12 AM, Evghenii Gaburov wrote:
Hi All,
Update to the original posting: METHOD4 also resulted in a deadlock on system HPC2 after
5h of run with 32 MPI tasks; also, "const int scale=1;" was missing in the code
snippet posted above.
--Evghenii
Message: 2
Date: Sun, 18 Sep 2011
I've not been following closely. How do you know you're using the
identical compilation flags? Are you saying you specify the same flags
to "mpicc" (or whatever) or are you confirming that the back-end
compiler is seeing the same flags? The MPI compiler wrapper (mpicc, et
al.) can add flags.
On 09/29/11 20:54, Xin Tong wrote:
I need to set up some environment variables before I run my
application ( appA ). I am currently using mpirun -np 1 -host socrates
(socrates is another machine) appA. Before appA runs, it expects
some environment variables to be set up. How do i do that ?
%
Often you set the environment variable DISPLAY to any display you like.
Export environment variables with "mpirun -x DISPLAY".
On 10/2/2011 5:32 AM, Xin Tong wrote:
I am launch a program with a GUI interface, How do i redirect the GUI
to the machine i issued mpirun on ?
Maybe someone else on this list has a better idea what you're trying to
do, but I'll attempt to answer your question.
MPI is basically a set of library calls that can be used for processes
to communicate with one another. Of course, a program need not have any
MPI calls in it, but if you want
I've not been following closely. Why must one use shared-memory
communications? How about using other BTLs in a "loopback" fashion?
something like --mca ^sm?
-Original Message-
From: users-boun...@open-mpi.org [mailto:users-boun...@open-mpi.org] On Behalf
Of Eugene Loh
Sent: Thursday, November 03, 2011 12:54 PM
To: us...@open-mpi.org
Subject: Re: [OMPI users] EXTERNAL: Re: How to set up state-less node /tmp for
Ope
On 11/10/2011 5:19 AM, Jeff Squyres wrote:
On Nov 10, 2011, at 8:11 AM, Mudassar Majeed wrote:
Thank you for your reply. I am implementing a load balancing function for MPI,
that will balance the computation load and the communication both at a time. So
my algorithm assumes that all the cores
If I remember correctly, both Intel MPI and MVAPICH2 bind processes by
default. OMPI does not. There are many cases where the "bind by
default" behavior gives better default performance. (There are also
cases where it can give catastrophically worse performance.) Anyhow, it
seems possible t
On 2/1/2012 7:59 AM, Frank wrote:
When running
mpirun -n 2
the STDOUT streams of both processes are combined and are displayed by
the shell. In such an interleaved format its hard to tell what line
comes from which node.
As far as this part goes, there is also "mpirun --tag-output". Check
the
Parallelize in distributed-memory fashion or is multi-threaded good
enough? Anyhow, you should be able to find many resources with an
Internet search. This particular mailing list is more for users of
OMPI, a particular MPI implementation. One approach would be to
distribute only one axis, s
On 03/13/12 13:25, Jeffrey Squyres wrote:
On Mar 9, 2012, at 5:17 PM, Jeremiah Willcock wrote:
On Open MPI 1.5.1, when I call MPI_Testsome with incount=0 and the two output
arrays NULL, I get an argument error (MPI_ERR_ARG). Is this the intended
behavior? If incount=0, no requests can comple
MPI_Wtime() returns the elapsed time since some arbitrary time in the
past. It is a measure of "wallclock" time, not of CPU time or anything.
On 5/4/2012 3:08 PM, Jingcha Joba wrote:
Lets say I have a code like this
start = MPI_Wtime()
stop = MPI_Wtime();
What happens when right after start=M
On 09/10/12 11:37, Ralph Castain wrote:
On Sep 10, 2012, at 8:12 AM, Aleksey Senin wrote:
On 10/09/2012 15:41, Siegmar Gross wrote:
Hi,
I have built openmpi-1.6.2rc1 and get the following error.
tyr small_prog 123 mpicc -showme
cc -I/usr/local/openmpi-1.6.2_32_cc/include -mt
-L/usr/local
On 11/5/2012 1:07 AM, huydanlin wrote:
Hi,
My objective is I want to calculate the time perform by MPI_Send &
MPI_Recv . In case MPI_Send, i can put the timer before the MPI_Send
and after its. like this "
t1=MPI_Wtime(),
MPI_Send
t2= MPI_Wtime
tsend= t2 -t1; it mean when the message go
On 02/05/13 00:30, Siegmar Gross wrote:
now I can use all our machines once more. I have a problem on
Solaris 10 x86_64, because the mapping of processes doesn't
correspond to the rankfile. I removed the output from "hostfile"
and wrapped around long lines.
tyr rankfiles 114 cat rf_ex_sunpc
# m
On 02/05/13 13:20, Eugene Loh wrote:
On 02/05/13 00:30, Siegmar Gross wrote:
now I can use all our machines once more. I have a problem on
Solaris 10 x86_64, because the mapping of processes doesn't
correspond to the rankfile.
A few comments.
First of all, the heterogeneous environmen
On 02/06/13 04:29, Siegmar Gross wrote:
Hi
thank you very much for your answer. I have compiled your program
and get different behaviours for openmpi-1.6.4rc3 and openmpi-1.9.
Yes, something else seems to be going on for 1.9.
For 1.6, try the attached patch. It works for me, but my machines
On 02/06/13 04:29, Siegmar Gross wrote:
thank you very much for your answer. I have compiled your program
and get different behaviours for openmpi-1.6.4rc3 and openmpi-1.9.
I get the following output for openmpi-1.9 (different outputs !!!).
sunpc1 rankfiles 104 mpirun --report-bindings --rankf
On 02/07/13 01:05, Siegmar Gross wrote:
thank you very much for your patch. I have applied the patch to
openmpi-1.6.4rc4.
Open MPI: 1.6.4rc4r28022
: [B .][. .] (slot list 0:0)
: [. B][. .] (slot list 0:1)
: [B B][. .] (slot list 0:0-1)
: [. .][B .] (slot list 1:0)
: [. .][. B] (slot list 1:1)
:
On 02/09/13 00:32, Ralph Castain wrote:
On Feb 6, 2013, at 2:59 PM, Eugene Loh wrote:
On 02/06/13 04:29, Siegmar Gross wrote:
thank you very much for your answer. I have compiled your program
and get different behaviours for openmpi-1.6.4rc3 and openmpi-1.9.
I think what's happening is
1 - 100 of 303 matches
Mail list logo