On Fri, Sep 3, 2010 at 3:47 PM, Jeff Squyres wrote:
> It might be worth having even a Linux-specific way to auto-detect, just for
> this use case (which is becoming more common -- 1GB LOM and a 10GB non-iWARP
> NIC).
The file:
/sys/class/net/ethX/speed
should contain the current speed and is
On Tue, Apr 27, 2010 at 7:55 PM, Samuel K. Gutierrez wrote:
> With Jeff and Ralph's help, I have completed a System V shared memory
> component for Open MPI.
What is the motivation for this work ? Are there situations where the
mmap based SM component doesn't work or is slow(er) ?
Kind regards,
On Sat, Apr 10, 2010 at 5:51 AM, Eugene Loh wrote:
> Why is shared-memory performance about four orders of magnitude slower than
> it should be?
Have there been any process scheduler changes in the newer kernels ?
I'm not sure that they could explain four orders of magnitude
differences though...
On Mon, Mar 1, 2010 at 9:15 PM, Ralph Castain wrote:
> Tracking this down has reminded me of all the reasons why I despise the
> rankfile mapper... :-/
Thanks for all your efforts ! I'm using the rankfile mapper as this is
the documented (in the FAQ) affinity-related one at least for the
stable
On Sat, Feb 27, 2010 at 7:35 PM, Ralph Castain wrote:
> I can't seem to replicate this first problem - it runs fine for me even if
> the rankfile contains only one entry.
First of all, thanks for taking a look at this !
For me it's repeatable. Please note that I do specify '-np 4' even
when in
Hi!
With version 1.4.1 I get a rather strange crash in mpirun whenever I
try to run a job using (I think) a rankfile which doesn't contain the
specified number of ranks. F.e. I ask for 4 ranks ('-np 4'), but the
rankfile contains only one entry:
rank 0=mbm-01-24 slot=1:*
and the following comes
ge the name of the package, f.e. to allow
installing several packages at the same time by simply changing:
Name: fftw2
to allow for the package called 'fftw' to track the 3.x versions. This
was done previously by Red Hat f.e. for their python packages.
--
Bogdan Costescu
IW
build should also solve the problem, or do I interpret things the
wrong way ?
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
several times. This, together with the earlier
post also describing a negative result, points to a problem related to
your particular setup...
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
e conditions and will write back in case this problem appears
again.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
8 in main ()
Can anyone suggest some ways forward ? I'd be happy to help in
debugging if given some instructions.
Thanks in advance!
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
ction "Network Support", there is a paragraph saying:
Open MPI will, by default, choose to use "cm" if it finds a
cm-supported network at run-time.
With the MX MTL being available at run-time, I would expect CM to be
chosen based on the quoted paragraph.
--
Bogdan Cos
be chosen; for v1.3rc3 I can't
distinguish anymore from timings as they behave very similarly.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
3rc3. That message has also raised the
question of selection of default components, but as there was no reply
to it, I still don't have any idea whether my testing or the docs were
wrong; can someone clear this up ?
Thanks for all the work put in v1.3 !
--
Bogdan Costescu
IWR, University of
I really unnecessary ? :-)
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
gs ? Can someone suggest what to do to
avoid them or at least a way to debug this ?
Thanks in advance !
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
to behave so poorly with MX ?
Thanks for any insight into this issues.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8240, Fax: +49 6221 54 8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
y available ? I'd be very much interested
in using it.
I've talked to George Bosilca at ISC08 about the issue of choosing the
right settings for (as close as possible to) maximum OpenMPI
performance on a given cluster and his answer was 'we regularly
organize workshops'
lf. This sounds very much
like the collectives tuning, with MCA params to give the admin or user
view of how the best performance can be achieved.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8869/8240, Fax: +49 6221 54 8868/8850
E-mai
On Tue, 12 Aug 2008, Rolf Vandevaart wrote:
I propose bumping the max for 32-bit programs to 2G and for 64-bit programs
to 8G.
Can't this be dynamically adjusted depending on the amount of RAM and
CPUs/cores ?
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heide
On Thu, 10 Jul 2008, Pavel Shamis (Pasha) wrote:
FYI the issue was resolved - https://svn.open-mpi.org/trac/ompi/ticket/1376
Indeed, no more IBCM error message displayed with r18878. Thank you !
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone
viour mentioned in a
previous e-mail (which happened with 1.3a1r18769) has disappeared,
CHARMM can again read its instructions properly from stdin.
Thanks for the quick resolution!
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8869/8240
setup your clusters have, but most that I have seen,
including all those that I admin, do run mpirun/mpiexec and rank=0 on
the same node. I really think that this will bite a lot of people.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54
t to it; my 1.3 list is already pretty long...
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8869/8240, Fax: +49 6221 54 8868/8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
l openib,self" so I think that the IB
stack is still being used (there is also a TCP/GigE network which
could be chosen otherwise).
I don't know whether this is caused by a somehow inconsistent setup of
the system, but I would welcome an option to make 1.3a behave like
1.2.
tested this recently with the RHEL5 kernels
with one gigabit and one Myri-10G connection, seeing a TCP stream
switching randomly between the gigabit and the Myri-10G connection.
--
Bogdan Costescu
IWR, University of Heidelberg, INF 368, D-69120 Heidelberg, Germany
Phone: +49 6221 54 8869/8240, Fax: +49 6221 54 8868/8850
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
On Tue, 30 Oct 2007, Bogdan Costescu wrote:
Bad timing... I don't have access to the files at the moment, I'll
write back shortly (which probably means tomorrow ;-)).
Here they are:
http://spider.iwr.uni-heidelberg.de/~bogdan/openmpi/
Due to their size, I decided to put them u
t the generated compile line for orterun
and it's library functions ended up being. This can include the
"libtool" lines.
Bad timing... I don't have access to the files at the moment, I'll
write back shortly (which probably means tomorrow ;-)).
--
Bogdan Costescu
ages already included the configure line; the
environment is only modified to set CC=pathcc, etc. and there are no
options on the make line; just like the documentation says that I
should do...
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Hei
and Brian recommend using
"--without-memory-manager", I won't feel bad about doing it :-)
Thanks a lot !
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 886
e compiler).
As I wrote in my previous e-mail, I tried configuring with and without
the MX libs, but this made no difference. It's only when I disabled
the memory manager, while still enabling MX, that I was able to get a
working build.
--
Bogdan Costescu
IWR - Interdisziplinaer
.2.3)
MCA sds: seed (MCA v1.0, API v1.0, Component v1.2.3)
MCA sds: singleton (MCA v1.0, API v1.0, Component v1.2.3)
MCA sds: pipe (MCA v1.0, API v1.0, Component v1.2.3)
MCA sds: slurm (MCA v1.0, API v1.0, Component v1
assing errors that I might have made, before filling a bug
report. The existing bugs related to PathScale compilers don't seem
to describe the symptoms that I'm seeing, unless it's some kind of
threading issue which seems to have no resolution yet...
Thanks in ad
ection and selection is
supposed to be for.
Yes, I understand that, it's the same type of mechanism as in LAM/MPI
which it's not that foreign to me ;-)
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120
case where there are several versions of the same batch
system installed, all using the same configuration files and therefore
being ready to run ? And how about the case where there is a machine
reserved for compilations, where libraries are made available but
there is no batch system active ?
y some effort on integrating ssh as well, the
problem being that the ssh daemon needs some modifications to allow
SGE to obtain accounting information. There was also some talk about a
TM-like API; unfortunately the progress in this area seems to be very
slow, if there is any at all...
--
Bo
ion is only about a proof of concept
version, then I'd say that anything to show IPv6 functionality would
be acceptable.
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54
munication via IPv4).
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
as IPs (either v4 or v6), OpenMPI should probably assume that
the address as given can be passed further to the underlying mechanism
for starting the job (for example, for SGE this would be its own rsh
client, not the system rsh client); but how about machines given as
names ?
--
Bo
ut ;-))
For example, we ran several weeks without an IPv6-enabled rsh, which
is used to handle MPI job startup on the cluster, without any
problems.
What do you mean by "IPv6-enabled rsh" ? Was it the daemon, client or
both ?
--
Bogdan Costescu
IWR - Interdisziplinaeres Zent
some user-provided
mapping.
That's all that I remember now from my IPv6 endeavour with LAM/MPI.
IMHO, some discussion of them should occur before the actual coding...
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 36
se changes
until a final, stable API was established, but when you want to be the
first to claim "I support this and that"...
Thanks for your diligence in pestering us about this! :-)
Eh, don't mention it! I want Open MPI to work :-)
--
Bogdan Costescu
IWR - Interdisziplinaeres
ssociated to case 1. of _get.
Cases 2. and 3. of _set are both associated to case 2. of _get.
So IMHO the test should be made with the _get function (as explained
in a previous message), by setting len=sizeof(long) which would allow
the case 1. to work fine, while case 2. would return -EINVAL, exac
ample by a (smart, don't know if such exists now) batch system.
I haven't checked if it's possible, but I think that a similar
solution based on sched_getaffinity would be much better, as this
would not disturb the current settings.
--
Bogdan Costescu
IWR - Interdiszipl
the _glibc_ function that
changes prototype.
--
Bogdan Costescu
IWR - Interdisziplinaeres Zentrum fuer Wissenschaftliches Rechnen
Universitaet Heidelberg, INF 368, D-69120 Heidelberg, GERMANY
Telephone: +49 6221 54 8869, Telefax: +49 6221 54 8868
E-mail: bogdan.coste...@iwr.uni-heidelberg.de
d, unsigned int, len, unsigned long
*, user_mask_ptr)
int main(int argc, char **argv){
unsigned long cpus = 1;
int r;
r = sched_setaffinity(0, sizeof(cpus), &cpus);
if (r == -1) {
perror("sched_setaffinity:");
}
return
46 matches
Mail list logo