[OMPI users] openmpi 1.2.1 on PK 2.6, and 64 bit version out
I've released a 64 bit version of parallelknoppix, a live CD that lets you set up a cluster for MPI-based parallel computing in about 10 minutes, without installing anything to the computers in the cluster. openmpi v1.2.1 (and lam and mpich) are all supported. Test reports are welcome, since this is an initial release and I have limited hardware to test it on. See www.parallelknoppix.net for more information.
Re: [OMPI users] Problem running hpcc with a threaded BLAS
On 4/27/07, Götz Waschk wrote: I'm testing my new cluster installation with the hpcc benchmark and openmpi 1.2.1 on RHEL5 32 bit. I have some trouble with using a threaded BLAS implementation. I have tried ATLAS 3.7.30 compiled with pthread support. It crashes as reported here: [...] I have a problem with Goto BLAS 1.14 too, the output of hpcc stops before the HPL run, then the hpcc processes seem to do nothing, consuming 100% CPU. If I set the maximum number of threads for Goto BLAS to 1, hpcc is working fine again. Hi, replying to myself here. I've tested this a bit more. It is working fine if I don't start hpcc from a Gridengine job. I think this is not related to openmpi's Gridengine integration, as the problem persists if I disable Gridengine integration on the mpirun command line. I'll keep you informed if I find a solution. Regards, Götz Waschk -- AL I:40: Do what thou wilt shall be the whole of the Law.
[OMPI users] torque, and mpiBlast
we use torque and tm for spawing mpi jobs on our cluster. We have a user who will be using mpiBlast. From their documentation they have: "Running a query on 25 nodes would look like: mpirun -np 27 mpiblast --config-file=/path/to/mpiblast.conf -p blastn -d nt -i blast_query.fas -o blast_results.txt" So they have 2 extra processes that are for special use that do not consume much cpu time thus they say spawn 2 more processes than dedicated. My question is how can I verify that the extra are placed on cpus that already have full working process (the other 25) on them? Is there a way to say place 25 process in the normal slots and rank 0 and 1, placed ware ever. Does this make sense? Brock Palen Center for Advanced Computing bro...@umich.edu (734)936-1985
