Re: [OMPI users] OpenMP & Open MPI

2016-03-13 Thread Jeff Hammond
On Sun, Mar 13, 2016 at 2:02 PM, Matthew Larkin wrote: > Hello, > > My understanding is Open MPI can utilize shared and/or distributed memory > architecture (parallel programming model). OpenMP is soley for shared > memory systems. > > The MPI-3 standard provides both explicitly shared and distri

Re: [OMPI users] OpenMP & Open MPI

2016-03-13 Thread Nick Papior
2016-03-13 23:04 GMT+01:00 Matthew Larkin : > Thanks Nick, > > I do understand they are different parallel paradigms and I understand the > purpose of each. > > If I run mpi processes to exhaust all cores, what is the need or purpose to > also use openmp to create threads? I believe there will be n

Re: [OMPI users] OpenMP & Open MPI

2016-03-13 Thread Matthew Larkin
Thanks Nick, I do understand they are different parallel paradigms and I understand the purpose of each.  If I run mpi processes to exhaust all cores, what is the need or purpose to also use openmp to create threads? I believe there will be no resources left for the forked threads?  Sent from Y

Re: [OMPI users] OpenMP & Open MPI

2016-03-13 Thread Nick Papior
2016-03-13 22:02 GMT+01:00 Matthew Larkin : > Hello, > > My understanding is Open MPI can utilize shared and/or distributed memory > architecture (parallel programming model). OpenMP is soley for shared memory > systems. > > I believe Open MPI incorporates OpenMP from some of the other archives I >

[OMPI users] OpenMP & Open MPI

2016-03-13 Thread Matthew Larkin
Hello, My understanding is Open MPI can utilize shared and/or distributed memory architecture (parallel programming model). OpenMP is soley for shared memory systems.  I believe Open MPI incorporates OpenMP from some of the other archives I glanced at.  Is this a true statement? If so, is there

Re: [OMPI users] Poor performance on Amazon EC2 with TCP

2016-03-13 Thread Jackson, Gary L.
By doing a parameter sweep, the best results I've gotten are with: btl_tcp_eager_limit & btl_tcp_rndv_eager_limit = 2 ** 17 btl_tcp_sndbuf & btl_tcp_rcvbuf = 2 ** 24 btl_tcp_endpoint_cache = 2 ** 12 btl_tcp_tcp_links = 2 Even so, the peak performance is around 7000Mbits/s and a message size aroun

Re: [OMPI users] Error with MPI_Register_datarep

2016-03-13 Thread George Bosilca
Eric, A quick grep in Open MPI source indicates that the only 2 places where MPI_ERR_UNSUPPORTED_DATAREP is issue are deep inside the imported ROMIO code (3.14): ./ompi/mca/io/romio314/romio/adio/include/adioi_errmsg.h:70:MPI_ERR_UNSUPPORTED_DATAREP ./ompi/mca/io/romio314/romio/adio/include/mpi