Further information.
I first time encountered this problem in openmpi-1.7.4.x,
while opnempi-1.7.3 and 1.6.x works fine.
My directory below is "testbed-openmpi-1.7.3", but it's realy
1.7.4a1r29646. I'm sorry, if I confuse you.
[mishima@manage testbed-openmpi-1.7.3]$ ompi_info | grep "Open MPI:
Indeed it should - most puzzling. I'll try playing with it on slurm using
sbatch and see if I get the same behavior. Offhand, I can't see why the
difference would exist unless somehow the script itself is taking one of the
execution slots, and somehow Torque is accounting for it.
Will have to e
Hi Ralph,
It's no problem that you let it lie until the problem becomes serious
again.
So, this is just an information for you.
I agree with your opinion that the problem will lie in the modified
hostfile.
But exactly speaking, it's related to just adding -hostfile option to
mpirun
in Torque s
On Nov 14, 2013, at 3:25 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Hi Ralph,
>
> I checked -cpus-per-proc in openmpi-1.7.4a1r29646.
> It works well as I want to do, which can adjust nprocs
> of each nodes dividing by number of threads.
>
> I think my problem is solved so far using -cpus-per
Hi Ralph,
I checked -cpus-per-proc in openmpi-1.7.4a1r29646.
It works well as I want to do, which can adjust nprocs
of each nodes dividing by number of threads.
I think my problem is solved so far using -cpus-per-proc,
thank you very mush.
Regarding oversbuscribed problem, I checked NPROCS was
FWIW: I verified that this works fine under a slurm allocation of 2 nodes, each
with 12 slots. I filled the node without getting an "oversbuscribed" error
message
[rhc@bend001 svn-trunk]$ mpirun -n 3 --bind-to core --cpus-per-proc 4
--report-bindings -hostfile hosts hostname
[bend001:24318] MCW
Also, you need to tell mpirun that the nodes aren't the same - add
--hetero-nodes to your cmd line
On Nov 13, 2013, at 10:14 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Thank you, Ralph!
>
> I didn't know that function of cups-per-proc.
> As fas as I know, it didn't work in openmpi-1.6.x lik
Thank you, Ralph!
I didn't know that function of cups-per-proc.
As fas as I know, it didn't work in openmpi-1.6.x like that.
It was just 4 cores binding...
Today I don't have much time and I'll check it tomorrow.
And thank you again for checking oversubscription problem.
tmishima
> Guess I d
Guess I don't see why modifying the allocation is required - we have mapping
options that should support such things. If you specify the total number of
procs you want, and cpus-per-proc=4, it should do the same thing I would think.
You'd get 2 procs on the 8 slot nodes, 8 on the 32 proc nodes,
Our cluster consists of three types of nodes. They have 8, 32
and 64 slots respectively. Since the performance of each core is
almost same, mixed use of these nodes is possible.
Furthremore, in this case, for hybrid application with openmpi+openmp,
the modification of hostfile is necesarry as fo
Why do it the hard way? I'll look at the FAQ because that definitely isn't a
recommended thing to do - better to use -host to specify the subset, or just
specify the desired mapping using all the various mappers we provide.
On Nov 13, 2013, at 6:39 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
>
Sorry for cross-post.
Nodefile is very simple which consists of 8 lines:
node08
node08
node08
node08
node08
node08
node08
node08
Therefore, NPROCS=8
My aim is to modify the allocation as you pointed out. According to Openmpi
FAQ,
proper subset of the hosts allocated to the Torque / PBS Pro jo
Sorry, I forgot to tell you.
Nodefile is very simple which consists of 8 lines:
node08
node08
node08
node08
node08
node08
node08
node08
tmishima
On Nov 13, 2013, at 4:43 PM, tmish...@jcity.maeda.co.jp wrote:
>
> >
> >
> > Yes, the node08 has 8 slots but the process I run is also 8.
> >
> > #
Please - can you answer my question on script2? What is the value of NPROCS?
Why would you want to do it this way? Are you planning to modify the
allocation?? That generally is a bad idea as it can confuse the system
On Nov 13, 2013, at 5:55 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Since
Since what I really want is to run script2 correctly, please let us
concentrate script2.
I'm not an expert of the inside of openmpi. What I can do is just
obsabation
from the outside. I doubt these lines are strange, especially the last one.
[node08.cluster:26952] mca:rmaps:rr: mapping job [565
On Nov 13, 2013, at 4:43 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Yes, the node08 has 8 slots but the process I run is also 8.
>
> #PBS -l nodes=node08:ppn=8
>
> Therefore, I think it should allow this allocation. Is that right?
Correct
>
> My question is why scritp1 works and script2 d
Yes, the node08 has 8 slots but the process I run is also 8.
#PBS -l nodes=node08:ppn=8
Therefore, I think it should allow this allocation. Is that right?
My question is why scritp1 works and script2 does not. They are
almost same.
#PBS -l nodes=node08:ppn=8
export OMP_NUM_THREADS=1
cd $PBS_
I guess here's my confusion. If you are using only one node, and that node has
8 allocated slots, then we will not allow you to run more than 8 processes on
that node unless you specifically provide the --oversubscribe flag. This is
because you are operating in a managed environment (in this cas
It has nothing to do with LAMA as you aren't using that mapper.
How many nodes are in this allocation?
On Nov 13, 2013, at 4:06 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Hi Ralph, this is an additional information.
>
> Here is the main part of output by adding "-mca rmaps_base_verbose 50".
Hi Ralph, this is an additional information.
Here is the main part of output by adding "-mca rmaps_base_verbose 50".
[node08.cluster:26952] [[56581,0],0] plm:base:setup_vm
[node08.cluster:26952] [[56581,0],0] plm:base:setup_vm creating map
[node08.cluster:26952] [[56581,0],0] plm:base:setup_vm
Hi Ralph,
This is the result of adding -mca ras_base_verbose 50.
SCRIPT:
mpirun -machinefile pbs_hosts -np ${NPROCS} -report-bindings -bind-to core
\
-mca ras_base_verbose 50 -mca plm_base_verbose 5 ./mPre
OUTPUT:
[node08.cluster:26770] mca:base:select:( plm) Querying component [rsh]
[
Hmmm...looks like we aren't getting your allocation. Can you rerun and add -mca
ras_base_verbose 50?
On Nov 12, 2013, at 11:30 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Hi Ralph,
>
> Here is the output of "-mca plm_base_verbose 5".
>
> [node08.cluster:23573] mca:base:select:( plm) Queryin
Hi Ralph,
Here is the output of "-mca plm_base_verbose 5".
[node08.cluster:23573] mca:base:select:( plm) Querying component [rsh]
[node08.cluster:23573] [[INVALID],INVALID] plm:rsh_lookup on
agent /usr/bin/rsh path NULL
[node08.cluster:23573] mca:base:select:( plm) Query of component [rsh] se
Hi Ralph,
Okey, I can help you. Please give me some time to report the output.
Tetsuya Mishima
> I can try, but I have no way of testing Torque any more - so all I can do
is a code review. If you can build --enable-debug and add -mca
plm_base_verbose 5 to your cmd line, I'd appreciate seeing t
I can try, but I have no way of testing Torque any more - so all I can do is a
code review. If you can build --enable-debug and add -mca plm_base_verbose 5 to
your cmd line, I'd appreciate seeing the output.
On Nov 12, 2013, at 9:58 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Hi Ralph,
>
>
Hi Ralph,
Thank you for your quick response.
I'd like to report one more regressive issue about Torque support of
openmpi-1.7.4a1r29646, which might be related to "#3893: LAMA mapper
has problems" I reported a few days ago.
The script below does not work with openmpi-1.7.4a1r29646,
although it
Done - thanks!
On Nov 12, 2013, at 7:35 PM, tmish...@jcity.maeda.co.jp wrote:
>
>
> Dear openmpi developers,
>
> I got a segmentation fault in traial use of openmpi-1.7.4a1r29646 built by
> PGI13.10 as shown below:
>
> [mishima@manage testbed-openmpi-1.7.3]$ mpirun -np 4 -cpus-per-proc 2
> -r
Dear openmpi developers,
I got a segmentation fault in traial use of openmpi-1.7.4a1r29646 built by
PGI13.10 as shown below:
[mishima@manage testbed-openmpi-1.7.3]$ mpirun -np 4 -cpus-per-proc 2
-report-bindings mPre
[manage.cluster:23082] MCW rank 2 bound to socket 0[core 4[hwt 0]], socket
0[c
28 matches
Mail list logo
