Open MPI logo

Open MPI Development Mailing List Archives

  |   Home   |   Support   |   FAQ   |   all Development mailing list

From: Matt Leininger (mlleinin_at_[hidden])
Date: 2006-11-28 00:06:18


On Mon, 2006-11-27 at 21:11 -0500, George Bosilca wrote:
> Which version of Open MPI are you using ? We can figure out what's
> wrong if we have the output of "ompi_info" and "ompi_info --param all
> all".

  Forgot the "ompi_info --param all all". It's attached.

  - Matt

>
> I wonder if some of the memory is not related to the size of the
> shared memory file. The default way to compute the size of the shared
> memory file is defined by the MCA parameter mpool_sm_per_peer_size.
> By default is set to 128MB for each local peer. Therefore using 2048
> procs on 256 nodes lead to using 8 procs by node i.e. at least 1GB
> only for the SM file. The problem right now with the SM file is that
> we're not reusing the buffers multiple times, instead we're using a
> new fragment each time we send a message, forcing the OS to map the
> entire file at one point.
>
> george.
>
> On Nov 27, 2006, at 8:21 PM, Matt Leininger wrote:
>
> > On Mon, 2006-11-27 at 16:45 -0800, Matt Leininger wrote:
> >> Has anyone testing OMPI's alltoall at > 2000 MPI tasks? I'm
> >> seeing each
> >> MPI task eat up > 1GB of memory (just for OMPI - not the app).
> >
> > I gathered some more data using the alltoall benchmark in mpiBench.
> > mpiBench is pretty smart about how large its buffers are. I set it to
> > use <= 100MB.
> >
> > num nodes num MPI tasks system mem mpibench buffer mem
> > 128 1024 1 GB 65 MB
> > 160 1280 1.2 GB 82 MB
> > 192 1536 1.4 GB 98 MB
> > 224 1792 1.6 GB 57 MB
> > 256 2048 1.6-1.8 GB < 100 MB
> >
> > The 256 node run was killed by the OOM for using too much memory. For
> > all these tests the OMPI alltoall is using 1 GB or more of system
> > memory. I know LANL is looking into optimized alltoall, but is anyone
> > looking into the scalability of the memory footprint?
> >
> > Thanks,
> >
> > - Matt
> >
> >>
> >> Thanks,
> >>
> >> - Matt
> >>
> >>
> >>
> >> _______________________________________________
> >> devel mailing list
> >> devel_at_[hidden]
> >> http://www.open-mpi.org/mailman/listinfo.cgi/devel
> >>
> >
> >
> > _______________________________________________
> > devel mailing list
> > devel_at_[hidden]
> > http://www.open-mpi.org/mailman/listinfo.cgi/devel
>
> _______________________________________________
> devel mailing list
> devel_at_[hidden]
> http://www.open-mpi.org/mailman/listinfo.cgi/devel
>


                 MCA mca: parameter "mca_param_files" (current value: "/g/g12/mlleinin/.openmpi/mca-params.conf:/g/g12/mlleinin/src/ompi-v1.2b-112506-gcc/etc/openmpi-mca-params.conf")
                          Path for MCA configuration files containing default parameter values
                 MCA mca: parameter "mca_component_path" (current value: "/g/g12/mlleinin/src/ompi-v1.2b-112506-gcc/lib/openmpi:/g/g12/mlleinin/.openmpi/components")
                          Path where to look for Open MPI and ORTE components
                 MCA mca: parameter "mca_verbose" (current value: <none>)
                          Top-level verbosity parameter
                 MCA mca: parameter "mca_component_show_load_errors" (current value: "1")
                          Whether to show errors for components that failed to load or not
                 MCA mca: parameter "mca_component_disable_dlopen" (current value: "0")
                          Whether to attempt to disable opening dynamic components or not
                 MCA mpi: parameter "mpi_param_check" (current value: "1")
                          Whether you want MPI API parameters checked at run-time or not. Possible values are 0 (no checking) and 1 (perform checking at run-time)
                 MCA mpi: parameter "mpi_yield_when_idle" (current value: "0")
                          Yield the processor when waiting for MPI communication (for MPI processes, will default to 1 when oversubscribing nodes)
                 MCA mpi: parameter "mpi_event_tick_rate" (current value: "-1")
                          How often to progress TCP communications (0 = never, otherwise specified in microseconds)
                 MCA mpi: parameter "mpi_show_handle_leaks" (current value: "0")
                          Whether MPI_FINALIZE shows all MPI handles that were not freed or not
                 MCA mpi: parameter "mpi_no_free_handles" (current value: "0")
                          Whether to actually free MPI objects when their handles are freed
                 MCA mpi: parameter "mpi_show_mca_params" (current value: "0")
                          Whether to show all MCA parameter value during MPI_INIT or not (good for reproducability of MPI jobs)
                 MCA mpi: parameter "mpi_show_mca_params_file" (current value: <none>)
                          If mpi_show_mca_params is true, setting this string to a valid filename tells Open MPI to dump all the MCA parameter values into a file suitable for reading via the mca_param_files parameter (good for reproducability of MPI jobs)
                 MCA mpi: parameter "mpi_paffinity_alone" (current value: "0")
                          If nonzero, assume that this job is the only (set of) process(es) running on each node and bind processes to processors, starting with processor ID 0
                 MCA mpi: parameter "mpi_keep_peer_hostnames" (current value: "1")
                          If nonzero, save the string hostnames of all MPI peer processes (mostly for error / debugging output messages). This can add quite a bit of memory usage to each MPI process.
                 MCA mpi: parameter "mpi_abort_delay" (current value: "0")
                          If nonzero, print out an identifying message when MPI_ABORT is invoked (hostname, PID of the process that called MPI_ABORT) and delay for that many seconds before exiting (a negative delay value means to never abort). This allows attaching of a debugger before quitting the job.
                 MCA mpi: parameter "mpi_abort_print_stack" (current value: "0")
                          If nonzero, print out a stack trace when MPI_ABORT is invoked
                 MCA mpi: parameter "mpi_preconnect_all" (current value: "0")
                          Whether to force MPI processes to create connections / warmup with *all* peers during MPI_INIT (vs. making connections lazily -- upon the first MPI traffic between each process peer pair)
                 MCA mpi: parameter "mpi_ddt_unpack_debug" (current value: "0")
                          Whether to output debugging information in the ddt unpack functions (nonzero = enabled)
                 MCA mpi: parameter "mpi_ddt_pack_debug" (current value: "0")
                          Whether to output debugging information in the ddt pack functions (nonzero = enabled)
                 MCA mpi: parameter "mpi_ddt_position_debug" (current value: "0")
                          Non zero lead to output generated by the datatype position functions
                 MCA mpi: parameter "mpi_ddt_copy_debug" (current value: "0")
                          Whether to output debugging information in the ddt copy functions (nonzero = enabled)
                 MCA mpi: parameter "mpi_leave_pinned" (current value: "0")
                          leave_pinned
                 MCA mpi: parameter "mpi_leave_pinned_pipeline" (current value: "0")
                          leave_pinned_pipeline
                MCA orte: parameter "orte_base_user_debugger" (current value: "totalview @mpirun@ -a @mpirun_args@ : fxp @mpirun@ -a @mpirun_args@")
                          Sequence of user-level debuggers to search for in orterun
                MCA orte: parameter "orte_debug" (current value: "0")
                          Whether or not to enable debugging output for all ORTE components (0 or 1)
                MCA orte: parameter "orte_debug_daemons" (current value: "0")
                          Whether or not to enable debugging of daemons (0 or 1)
                MCA orte: parameter "orte_timing" (current value: "0")
                          Request that critical timing loops be measured
                MCA opal: parameter "opal_signal" (current value: "6,7,8,11")
                          If a signal is received, display the stack trace frame
           MCA backtrace: parameter "backtrace" (current value: <none>)
                          Default selection set of components for the backtrace framework (<none> means "use all components that can be found")
           MCA backtrace: parameter "backtrace_base_verbose" (current value: "0")
                          Verbosity level for the backtrace framework (0 = no verbosity)
           MCA backtrace: parameter "backtrace_execinfo_priority" (current value: "0")
              MCA memory: parameter "memory" (current value: <none>)
                          Default selection set of components for the memory framework (<none> means "use all components that can be found")
              MCA memory: parameter "memory_base_verbose" (current value: "0")
                          Verbosity level for the memory framework (0 = no verbosity)
              MCA memory: parameter "memory_ptmalloc2_priority" (current value: "0")
           MCA paffinity: parameter "paffinity" (current value: <none>)
                          Default selection set of components for the paffinity framework (<none> means "use all components that can be found")
           MCA paffinity: parameter "paffinity_linux_priority" (current value: "10")
                          Priority of the linux paffinity component
           MCA paffinity: information "paffinity_linux_have_cpu_set_t" (value: "1")
                          Whether this component was compiled on a system with the type cpu_set_t or not (1 = yes, 0 = no)
           MCA paffinity: information "paffinity_linux_CPU_ZERO_ok" (value: "1")
                          Whether this component was compiled on a system where CPU_ZERO() is functional or broken (1 = functional, 0 = broken/not available)
           MCA paffinity: information "paffinity_linux_sched_setaffinity_num_params" (value: "3")
                          The number of parameters that sched_set_affinity() takes on the machine where this component was compiled
           MCA maffinity: parameter "maffinity" (current value: <none>)
                          Default selection set of components for the maffinity framework (<none> means "use all components that can be found")
           MCA maffinity: parameter "maffinity_first_use_priority" (current value: "10")
                          Priority of the first_use maffinity component
               MCA timer: parameter "timer" (current value: <none>)
                          Default selection set of components for the timer framework (<none> means "use all components that can be found")
               MCA timer: parameter "timer_base_verbose" (current value: "0")
                          Verbosity level for the timer framework (0 = no verbosity)
               MCA timer: parameter "timer_linux_priority" (current value: "0")
           MCA allocator: parameter "allocator" (current value: <none>)
                          Default selection set of components for the allocator framework (<none> means "use all components that can be found")
           MCA allocator: parameter "allocator_base_verbose" (current value: "0")
                          Verbosity level for the allocator framework (0 = no verbosity)
           MCA allocator: parameter "allocator_basic_priority" (current value: "0")
           MCA allocator: parameter "allocator_bucket_num_buckets" (current value: "30")
           MCA allocator: parameter "allocator_bucket_priority" (current value: "0")
                MCA coll: parameter "coll" (current value: <none>)
                          Default selection set of components for the coll framework (<none> means "use all components that can be found")
                MCA coll: parameter "coll_base_verbose" (current value: "0")
                          Verbosity level for the coll framework (0 = no verbosity)
                MCA coll: parameter "coll_basic_priority" (current value: "10")
                          Priority of the basic coll component
                MCA coll: parameter "coll_basic_crossover" (current value: "4")
                          Minimum number of processes in a communicator before using the logarithmic algorithms
                MCA coll: parameter "coll_self_priority" (current value: "75")
                MCA coll: parameter "coll_sm_priority" (current value: "0")
                          Priority of the sm coll component
                MCA coll: parameter "coll_sm_control_size" (current value: "4096")
                          Length of the control data -- should usually be either the length of a cache line on most SMPs, or the size of a page on machines that support direct memory affinity page placement (in bytes)
                MCA coll: parameter "coll_sm_bootstrap_filename" (current value: "shared_mem_sm_bootstrap")
                          Filename (in the Open MPI session directory) of the coll sm component bootstrap rendezvous mmap file
                MCA coll: parameter "coll_sm_bootstrap_num_segments" (current value: "8")
                          Number of segments in the bootstrap file
                MCA coll: parameter "coll_sm_fragment_size" (current value: "8192")
                          Fragment size (in bytes) used for passing data through shared memory (will be rounded up to the nearest control_size size)
                MCA coll: parameter "coll_sm_mpool" (current value: "sm")
                          Name of the mpool component to use
                MCA coll: parameter "coll_sm_comm_in_use_flags" (current value: "2")
                          Number of "in use" flags, used to mark a message passing area segment as currently being used or not (must be >= 2 and <= comm_num_segments)
                MCA coll: parameter "coll_sm_comm_num_segments" (current value: "8")
                          Number of segments in each communicator's shared memory message passing area (must be >= 2, and must be a multiple of comm_in_use_flags)
                MCA coll: parameter "coll_sm_tree_degree" (current value: "4")
                          Degree of the tree for tree-based operations (must be => 1 and <= min(control_size, 255))
                MCA coll: information "coll_sm_shared_mem_used_bootstrap" (value: "216")
                          Amount of shared memory used in the shared memory bootstrap area (in bytes)
                MCA coll: parameter "coll_sm_info_num_procs" (current value: "4")
                          Number of processes to use for the calculation of the shared_mem_size MCA information parameter (must be => 2)
                MCA coll: information "coll_sm_shared_mem_used_data" (value: "548864")
                          Amount of shared memory used in the shared memory data area for info_num_procs processes (in bytes)
                MCA coll: parameter "coll_tuned_priority" (current value: "30")
                          Priority of the tuned coll component
                MCA coll: parameter "coll_tuned_pre_allocate_memory_comm_size_limit" (current value: "32768")
                          Size of communicator were we stop pre-allocating memory for the fixed internal buffer used for message requests etc that is hung off the communicator data segment. I.e. if you have a 100'000 nodes you might not want to pre-allocate 200'000 request handle slots per communicator instance!
                MCA coll: parameter "coll_tuned_use_dynamic_rules" (current value: "0")
                          Switch used to decide if we use static (compiled/if statements) or dynamic (built at runtime) decision function rules
                MCA coll: parameter "coll_tuned_init_tree_fanout" (current value: "4")
                          Inital fanout used in the tree topologies for each communicator. This is only an initial guess, if a tuned collective needs a different fanout for an operation, it build it dynamically. This parameter is only for the first guess and might save a little time
                MCA coll: parameter "coll_tuned_init_chain_fanout" (current value: "4")
                          Inital fanout used in the chain (fanout followed by pipeline) topologies for each communicator. This is only an initial guess, if a tuned collective needs a different fanout for an operation, it build it dynamically. This parameter is only for the first guess and might save a little time
                  MCA io: parameter "io_base_freelist_initial_size" (current value: "16")
                          Initial MPI-2 IO request freelist size
                  MCA io: parameter "io_base_freelist_max_size" (current value: "64")
                          Max size of the MPI-2 IO request freelist
                  MCA io: parameter "io_base_freelist_increment" (current value: "16")
                          Increment size of the MPI-2 IO request freelist
                  MCA io: parameter "io" (current value: <none>)
                          Default selection set of components for the io framework (<none> means "use all components that can be found")
                  MCA io: parameter "io_base_verbose" (current value: "0")
                          Verbosity level for the io framework (0 = no verbosity)
                  MCA io: parameter "io_romio_priority" (current value: "10")
                          Priority of the io romio component
                  MCA io: parameter "io_romio_delete_priority" (current value: "10")
                          Delete priority of the io romio component
                  MCA io: parameter "io_romio_enable_parallel_optimizations" (current value: "0")
                          Enable set of Open MPI-added options to improve collective file i/o performance
               MCA mpool: parameter "mpool" (current value: <none>)
                          Default selection set of components for the mpool framework (<none> means "use all components that can be found")
               MCA mpool: parameter "mpool_base_verbose" (current value: "0")
                          Verbosity level for the mpool framework (0 = no verbosity)
               MCA mpool: parameter "mpool_openib_rcache_name" (current value: "rb")
                          The name of the registration cache the mpool should use
               MCA mpool: parameter "mpool_openib_priority" (current value: "0")
               MCA mpool: parameter "mpool_sm_allocator" (current value: "bucket")
                          Name of allocator component to use with sm mpool
               MCA mpool: parameter "mpool_sm_max_size" (current value: "536870912")
                          Maximum size of the sm mpool shared memory file
               MCA mpool: parameter "mpool_sm_min_size" (current value: "134217728")
                          Minimum size of the sm mpool shared memory file
               MCA mpool: parameter "mpool_sm_per_peer_size" (current value: "33554432")
                          Size (in bytes) to allocate per local peer in the sm mpool shared memory file, bounded by min_size and max_size
               MCA mpool: parameter "mpool_sm_priority" (current value: "0")
               MCA mpool: parameter "mpool_udapl_priority" (current value: "0")
               MCA mpool: parameter "mpool_base_use_mem_hooks" (current value: "0")
                          use memory hooks for deregistering freed memory
               MCA mpool: parameter "mpool_use_mem_hooks" (current value: "0")
                          (deprecated, use mpool_base_use_mem_hooks)
                 MCA pml: parameter "pml" (current value: "ob1")
                          Default selection set of components for the pml framework (<none> means "use all components that can be found")
                 MCA pml: parameter "pml_base_verbose" (current value: "0")
                          Verbosity level for the pml framework (0 = no verbosity)
                 MCA pml: parameter "pml_cm_free_list_num" (current value: "4")
                          Initial size of request free lists
                 MCA pml: parameter "pml_cm_free_list_max" (current value: "-1")
                          Maximum size of request free lists
                 MCA pml: parameter "pml_cm_free_list_inc" (current value: "64")
                          Number of elements to add when growing request free lists
                 MCA pml: parameter "pml_cm_priority" (current value: "1")
                          CM PML selection priority
                 MCA pml: parameter "pml_dr_free_list_num" (current value: "4")
                 MCA pml: parameter "pml_dr_free_list_max" (current value: "-1")
                 MCA pml: parameter "pml_dr_free_list_inc" (current value: "64")
                 MCA pml: parameter "pml_dr_priority" (current value: "1")
                 MCA pml: parameter "pml_dr_eager_limit" (current value: "131072")
                 MCA pml: parameter "pml_dr_send_pipeline_depth" (current value: "3")
                 MCA pml: parameter "pml_dr_wdog_timer_sec" (current value: "5")
                 MCA pml: parameter "pml_dr_wdog_timer_usec" (current value: "0")
                 MCA pml: parameter "pml_dr_wdog_timer_multiplier" (current value: "1")
                 MCA pml: parameter "pml_dr_wdog_retry_max" (current value: "1")
                 MCA pml: parameter "pml_dr_ack_timer_sec" (current value: "10")
                 MCA pml: parameter "pml_dr_ack_timer_usec" (current value: "0")
                 MCA pml: parameter "pml_dr_ack_timer_multiplier" (current value: "1")
                 MCA pml: parameter "pml_dr_ack_retry_max" (current value: "3")
                 MCA pml: parameter "pml_dr_enable_csum" (current value: "1")
                 MCA pml: parameter "pml_ob1_free_list_num" (current value: "4")
                 MCA pml: parameter "pml_ob1_free_list_max" (current value: "-1")
                 MCA pml: parameter "pml_ob1_free_list_inc" (current value: "64")
                 MCA pml: parameter "pml_ob1_priority" (current value: "1")
                 MCA pml: parameter "pml_ob1_eager_limit" (current value: "131072")
                 MCA pml: parameter "pml_ob1_send_pipeline_depth" (current value: "3")
                 MCA pml: parameter "pml_ob1_recv_pipeline_depth" (current value: "4")
                 MCA bml: parameter "bml" (current value: <none>)
                          Default selection set of components for the bml framework (<none> means "use all components that can be found")
                 MCA bml: parameter "bml_base_verbose" (current value: "0")
                          Verbosity level for the bml framework (0 = no verbosity)
                 MCA bml: parameter "bml_r2_show_unreach_errors" (current value: "1")
                          Show error message when procs are unreachable
                 MCA bml: parameter "bml_r2_priority" (current value: "0")
              MCA rcache: parameter "rcache" (current value: <none>)
                          Default selection set of components for the rcache framework (<none> means "use all components that can be found")
              MCA rcache: parameter "rcache_base_verbose" (current value: "0")
                          Verbosity level for the rcache framework (0 = no verbosity)
              MCA rcache: parameter "rcache_rb_priority" (current value: "0")
              MCA rcache: parameter "rcache_vma_mru_len" (current value: "256")
                          The maximum size IN ENTRIES of the MRU (most recently used) rcache list
              MCA rcache: parameter "rcache_vma_mru_size" (current value: "1073741824")
                          The maximum size IN BYTES of the MRU (most recently used) rcache list
              MCA rcache: parameter "rcache_vma_priority" (current value: "0")
                 MCA btl: parameter "btl_base_debug" (current value: "0")
                          If btl_base_debug is 1 standard debug is output, if > 1 verbose debug is output
                 MCA btl: parameter "btl" (current value: <none>)
                          Default selection set of components for the btl framework (<none> means "use all components that can be found")
                 MCA btl: parameter "btl_base_verbose" (current value: "0")
                          Verbosity level for the btl framework (0 = no verbosity)
                 MCA btl: parameter "btl_openib_verbose" (current value: "0")
                          Output some verbose OpenIB BTL information (0 = no output, nonzero = output)
                 MCA btl: parameter "btl_openib_warn_no_hca_params_found" (current value: "1")
                          Warn when no HCA-specific parameters are found in the INI file specified by the btl_openib_hca_param_files MCA parameter (0 = do not warn; any other value = warn)
                 MCA btl: parameter "btl_openib_warn_default_gid_prefix" (current value: "1")
                          Warn when there is more than one active ports and at least one of them connected to the network with only default GID prefix configured (0 = do not warn; any other value = warn)
                 MCA btl: parameter "btl_openib_hca_param_files" (current value: "/g/g12/mlleinin/src/ompi-v1.2b-112506-gcc/share/openmpi/mca-btl-openib-hca-params.ini")
                          Colon-delimited list of INI-style files that contain HCA vendor/part-specific parameters
                 MCA btl: parameter "btl_openib_max_btls" (current value: "-1")
                          Maximum number of HCA ports to use (-1 = use all available, otherwise must be >= 1)
                 MCA btl: parameter "btl_openib_free_list_num" (current value: "8")
                          Intial size of free lists (must be >= 1)
                 MCA btl: parameter "btl_openib_free_list_max" (current value: "-1")
                          Maximum size of free lists (-1 = infinite, otherwise must be >= 0)
                 MCA btl: parameter "btl_openib_free_list_inc" (current value: "32")
                          Increment size of free lists (must be >= 1)
                 MCA btl: parameter "btl_openib_mpool" (current value: "openib")
                          Name of the memory pool to be used (it is unlikely that you will ever want to change this
                 MCA btl: parameter "btl_openib_reg_mru_len" (current value: "16")
                          Length of the registration cache most recently used list (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_cq_size" (current value: "1000")
                          Size of the IB completion queue (will automatically be set to a minimum of (2 * number_of_peers * btl_openib_rd_num))
                 MCA btl: parameter "btl_openib_ib_sg_list_size" (current value: "4")
                          Size of IB segment list (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_pkey_ix" (current value: "0")
                          InfiniBand pkey index (must be >= 0)
                 MCA btl: parameter "btl_openib_ib_psn" (current value: "0")
                          InfiniBand packet sequence starting number (must be >= 0)
                 MCA btl: parameter "btl_openib_ib_qp_ous_rd_atom" (current value: "4")
                          InfiniBand outstanding atomic reads (must be >= 0)
                 MCA btl: parameter "btl_openib_ib_mtu" (current value: "3")
                          IB MTU, in bytes (if not specified in INI files). Valid values are: 1=256 bytes, 2=512 bytes, 3=1024 bytes, 4=2048 bytes, 5=4096 bytes
                 MCA btl: parameter "btl_openib_ib_min_rnr_timer" (current value: "5")
                          InfiniBand minimum "receiver not ready" timer, in seconds (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_timeout" (current value: "10")
                          InfiniBand transmit timeout, in seconds(must be >= 1)
                 MCA btl: parameter "btl_openib_ib_retry_count" (current value: "7")
                          InfiniBand transmit retry count (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_rnr_retry" (current value: "7")
                          InfiniBand "receiver not ready" retry count (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_max_rdma_dst_ops" (current value: "4")
                          InfiniBand maximum pending RDMA destination operations (must be >= 1)
                 MCA btl: parameter "btl_openib_ib_service_level" (current value: "0")
                          InfiniBand service level (must be >= 0)
                 MCA btl: parameter "btl_openib_ib_static_rate" (current value: "0")
                          InfiniBand static rate (must be >= 0; defulat: %d)
                 MCA btl: parameter "btl_openib_exclusivity" (current value: "1024")
                          OpenIB BTL exclusivity (must be >= 0)
                 MCA btl: parameter "btl_openib_rd_num" (current value: "8")
                          Number of receive descriptors to post to a queue pair (must be >= 1)
                 MCA btl: parameter "btl_openib_rd_low" (current value: "6")
                          Low water mark before reposting occurs (must be >= 1)
                 MCA btl: parameter "btl_openib_rd_win" (current value: "4")
                          Window size at which generate explicit credit message (must be >= 1)
                 MCA btl: parameter "btl_openib_use_srq" (current value: "0")
                          If nonzero, use the InfiniBand shared receive queue ("SRQ")
                 MCA btl: parameter "btl_openib_srq_rd_max" (current value: "1000")
                          Maxium number of receive descriptors posted per SRQ (only relevant if btl_openib_use_srq is true; must be >= 1)
                 MCA btl: parameter "btl_openib_srq_rd_per_peer" (current value: "16")
                          Number of receive descriptors posted per peer in the SRQ (only relevant if btl_openib_use_srq is true; must be >= 1)
                 MCA btl: parameter "btl_openib_srq_sd_max" (current value: "8")
                          Maximum number of send descriptors posted (only relevant if btl_openib_use_srq is true; must be >= 1)
                 MCA btl: parameter "btl_openib_use_eager_rdma" (current value: "1")
                          Use RDMA for eager messages
                 MCA btl: parameter "btl_openib_eager_rdma_threshold" (current value: "16")
                          Use RDMA for short messages after this number of messages are received from a given peer (must be >= 1)
                 MCA btl: parameter "btl_openib_max_eager_rdma" (current value: "16")
                          Maximum number of peers allowed to use RDMA for short messages (RDMA is used for all long messages, except if explicitly disabled, such as with the "dr" pml) (must be >= 0)
                 MCA btl: parameter "btl_openib_eager_rdma_num" (current value: "16")
                          Number of RDMA buffers to allocate for small messages(must be >= 1)
                 MCA btl: parameter "btl_openib_btls_per_lid" (current value: "1")
                          Number of BTLs to create for each InfiniBand LID (must be >= 1)
                 MCA btl: parameter "btl_openib_max_lmc" (current value: "0")
                          Maximum number of LIDs to use for each HCA port (must be >= 0, where 0 = use all available)
                 MCA btl: parameter "btl_openib_buffer_alignment" (current value: "64")
                          Prefered communication buffer alignment, in bytes (must be >= 0)
                 MCA btl: parameter "btl_openib_eager_limit" (current value: "12288")
                          Eager send limit, in bytes (must be >= 1)
                 MCA btl: parameter "btl_openib_min_send_size" (current value: "32768")
                          Minimum send size, in bytes (must be >= 1)
                 MCA btl: parameter "btl_openib_max_send_size" (current value: "65536")
                          Maximum send size, in bytes (must be >= 1)
                 MCA btl: parameter "btl_openib_min_rdma_size" (current value: "1048576")
                          Minimum RDMA size, in bytes (must be >= 1)
                 MCA btl: parameter "btl_openib_max_rdma_size" (current value: "1048576")
                          Maximium RDMA size, in bytes (must be >= 1)
                 MCA btl: parameter "btl_openib_flags" (current value: "54")
                          BTL flags, added together: SEND=1, PUT=2, GET=4 (cannot be 0)
                 MCA btl: parameter "btl_openib_bandwidth" (current value: "800")
                          Approximate maximum bandwidth of network (must be >= 1)
                 MCA btl: parameter "btl_openib_priority" (current value: "0")
                 MCA btl: parameter "btl_self_free_list_num" (current value: "0")
                          Number of fragments by default
                 MCA btl: parameter "btl_self_free_list_max" (current value: "-1")
                          Maximum number of fragments
                 MCA btl: parameter "btl_self_free_list_inc" (current value: "32")
                          Increment by this number of fragments
                 MCA btl: parameter "btl_self_eager_limit" (current value: "131072")
                          Eager size fragmeng (before the rendez-vous ptotocol)
                 MCA btl: parameter "btl_self_min_send_size" (current value: "262144")
                          Minimum fragment size after the rendez-vous
                 MCA btl: parameter "btl_self_max_send_size" (current value: "262144")
                          Maximum fragment size after the rendez-vous
                 MCA btl: parameter "btl_self_min_rdma_size" (current value: "2147483647")
                          Maximum fragment size for the RDMA transfer
                 MCA btl: parameter "btl_self_max_rdma_size" (current value: "2147483647")
                          Maximum fragment size for the RDMA transfer
                 MCA btl: parameter "btl_self_exclusivity" (current value: "65536")
                          Device exclusivity
                 MCA btl: parameter "btl_self_flags" (current value: "10")
                          Active behavior flags
                 MCA btl: parameter "btl_self_priority" (current value: "0")
                 MCA btl: parameter "btl_sm_free_list_num" (current value: "8")
                 MCA btl: parameter "btl_sm_free_list_max" (current value: "-1")
                 MCA btl: parameter "btl_sm_free_list_inc" (current value: "64")
                 MCA btl: parameter "btl_sm_exclusivity" (current value: "65535")
                 MCA btl: parameter "btl_sm_latency" (current value: "100")
                 MCA btl: parameter "btl_sm_max_procs" (current value: "-1")
                 MCA btl: parameter "btl_sm_sm_extra_procs" (current value: "2")
                 MCA btl: parameter "btl_sm_mpool" (current value: "sm")
                 MCA btl: parameter "btl_sm_eager_limit" (current value: "4096")
                 MCA btl: parameter "btl_sm_max_frag_size" (current value: "32768")
                 MCA btl: parameter "btl_sm_size_of_cb_queue" (current value: "128")
                 MCA btl: parameter "btl_sm_cb_lazy_free_freq" (current value: "120")
                 MCA btl: parameter "btl_sm_priority" (current value: "0")
                 MCA btl: parameter "btl_tcp_if_include" (current value: <none>)
                 MCA btl: parameter "btl_tcp_if_exclude" (current value: "lo")
                 MCA btl: parameter "btl_tcp_free_list_num" (current value: "8")
                 MCA btl: parameter "btl_tcp_free_list_max" (current value: "-1")
                 MCA btl: parameter "btl_tcp_free_list_inc" (current value: "32")
                 MCA btl: parameter "btl_tcp_sndbuf" (current value: "131072")
                 MCA btl: parameter "btl_tcp_rcvbuf" (current value: "131072")
                 MCA btl: parameter "btl_tcp_endpoint_cache" (current value: "30720")
                 MCA btl: parameter "btl_tcp_exclusivity" (current value: "0")
                 MCA btl: parameter "btl_tcp_eager_limit" (current value: "65536")
                 MCA btl: parameter "btl_tcp_min_send_size" (current value: "65536")
                 MCA btl: parameter "btl_tcp_max_send_size" (current value: "131072")
                 MCA btl: parameter "btl_tcp_min_rdma_size" (current value: "131072")
                 MCA btl: parameter "btl_tcp_max_rdma_size" (current value: "2147483647")
                 MCA btl: parameter "btl_tcp_flags" (current value: "58")
                 MCA btl: parameter "btl_tcp_priority" (current value: "0")
                 MCA btl: parameter "btl_udapl_free_list_num" (current value: "8")
                 MCA btl: parameter "btl_udapl_free_list_max" (current value: "-1")
                 MCA btl: parameter "btl_udapl_free_list_inc" (current value: "8")
                 MCA btl: parameter "btl_udapl_mpool" (current value: "udapl")
                 MCA btl: parameter "btl_udapl_max_modules" (current value: "8")
                 MCA btl: parameter "btl_udapl_evd_qlen" (current value: "32")
                 MCA btl: parameter "btl_udapl_num_recvs" (current value: "8")
                 MCA btl: parameter "btl_udapl_num_sends" (current value: "8")
                 MCA btl: parameter "btl_udapl_timeout" (current value: "10000000")
                 MCA btl: parameter "btl_udapl_exclusivity" (current value: "1014")
                 MCA btl: parameter "btl_udapl_eager_limit" (current value: "32768")
                 MCA btl: parameter "btl_udapl_min_send_size" (current value: "16384")
                 MCA btl: parameter "btl_udapl_max_send_size" (current value: "65536")
                 MCA btl: parameter "btl_udapl_min_rdma_size" (current value: "524288")
                 MCA btl: parameter "btl_udapl_max_rdma_size" (current value: "131072")
                 MCA btl: parameter "btl_udapl_bandwidth" (current value: "225")
                 MCA btl: parameter "btl_udapl_priority" (current value: "0")
                 MCA btl: parameter "btl_base_include" (current value: <none>)
                 MCA btl: parameter "btl_base_exclude" (current value: <none>)
                 MCA btl: parameter "btl_base_warn_component_unused" (current value: "1")
                          This parameter is used to turn on warning messages when certain NICs are not used
                 MCA mtl: parameter "mtl" (current value: <none>)
                          Default selection set of components for the mtl framework (<none> means "use all components that can be found")
                 MCA mtl: parameter "mtl_base_verbose" (current value: "0")
                          Verbosity level for the mtl framework (0 = no verbosity)
                MCA topo: parameter "topo" (current value: <none>)
                          Default selection set of components for the topo framework (<none> means "use all components that can be found")
                MCA topo: parameter "topo_base_verbose" (current value: "0")
                          Verbosity level for the topo framework (0 = no verbosity)
                 MCA osc: parameter "osc" (current value: <none>)
                          Default selection set of components for the osc framework (<none> means "use all components that can be found")
                 MCA osc: parameter "osc_base_verbose" (current value: "0")
                          Verbosity level for the osc framework (0 = no verbosity)
                 MCA osc: parameter "osc_pt2pt_no_locks" (current value: "0")
                          Enable optimizations available only if MPI_LOCK is not used.
                 MCA osc: parameter "osc_pt2pt_eager_limit" (current value: "16384")
                          Max size of eagerly sent data
                 MCA osc: parameter "osc_pt2pt_priority" (current value: "0")
                 MCA osc: parameter "osc_rdma_fence_sync_method" (current value: "reduce_scatter")
                          How to synchronize fence: reduce_scatter, allreduce, alltoall
                 MCA osc: parameter "osc_rdma_eager_send" (current value: "0")
                          Attempt to start data movement during communication call, instead of at synchrnoization time. Info key of same name overrides this value, if info key given.
                 MCA osc: parameter "osc_rdma_no_locks" (current value: "0")
                          Enable optimizations available only if MPI_LOCK is not used.
                 MCA osc: parameter "osc_rdma_priority" (current value: "0")
              MCA errmgr: parameter "errmgr" (current value: <none>)
                          Default selection set of components for the errmgr framework (<none> means "use all components that can be found")
              MCA errmgr: parameter "errmgr_hnp_debug" (current value: "0")
              MCA errmgr: parameter "errmgr_hnp_priority" (current value: "0")
              MCA errmgr: parameter "errmgr_orted_debug" (current value: "0")
              MCA errmgr: parameter "errmgr_orted_priority" (current value: "0")
              MCA errmgr: parameter "errmgr_proxy_debug" (current value: "0")
              MCA errmgr: parameter "errmgr_proxy_priority" (current value: "0")
                 MCA gpr: parameter "gpr_base_maxsize" (current value: "2147483647")
                 MCA gpr: parameter "gpr_base_blocksize" (current value: "512")
                 MCA gpr: parameter "gpr" (current value: <none>)
                          Default selection set of components for the gpr framework (<none> means "use all components that can be found")
                 MCA gpr: parameter "gpr_null_priority" (current value: "0")
                 MCA gpr: parameter "gpr_proxy_debug" (current value: "0")
                 MCA gpr: parameter "gpr_proxy_priority" (current value: "0")
                 MCA gpr: parameter "gpr_replica_debug" (current value: "0")
                 MCA gpr: parameter "gpr_replica_isolate" (current value: "0")
                 MCA gpr: parameter "gpr_replica_priority" (current value: "0")
                 MCA iof: parameter "iof_base_window_size" (current value: "4096")
                 MCA iof: parameter "iof_base_service" (current value: "0.0.0")
                 MCA iof: parameter "iof" (current value: <none>)
                          Default selection set of components for the iof framework (<none> means "use all components that can be found")
                 MCA iof: parameter "iof_proxy_debug" (current value: "1")
                 MCA iof: parameter "iof_proxy_priority" (current value: "0")
                 MCA iof: parameter "iof_svc_debug" (current value: "1")
                 MCA iof: parameter "iof_svc_priority" (current value: "0")
                  MCA ns: parameter "ns" (current value: <none>)
                          Default selection set of components for the ns framework (<none> means "use all components that can be found")
                  MCA ns: parameter "ns_proxy_debug" (current value: "0")
                  MCA ns: parameter "ns_proxy_maxsize" (current value: "2147483647")
                  MCA ns: parameter "ns_proxy_blocksize" (current value: "512")
                  MCA ns: parameter "ns_proxy_priority" (current value: "0")
                  MCA ns: parameter "ns_replica_debug" (current value: "0")
                  MCA ns: parameter "ns_replica_isolate" (current value: "0")
                  MCA ns: parameter "ns_replica_maxsize" (current value: "2147483647")
                  MCA ns: parameter "ns_replica_blocksize" (current value: "512")
                  MCA ns: parameter "ns_replica_priority" (current value: "0")
                 MCA oob: parameter "oob" (current value: <none>)
                          Default selection set of components for the oob framework (<none> means "use all components that can be found")
                 MCA oob: parameter "oob_base_verbose" (current value: "0")
                          Verbosity level for the oob framework (0 = no verbosity)
                 MCA oob: parameter "oob_tcp_peer_limit" (current value: "-1")
                 MCA oob: parameter "oob_tcp_peer_retries" (current value: "60")
                 MCA oob: parameter "oob_tcp_debug" (current value: "0")
                 MCA oob: parameter "oob_tcp_include" (current value: <none>)
                 MCA oob: parameter "oob_tcp_exclude" (current value: <none>)
                 MCA oob: parameter "oob_tcp_sndbuf" (current value: "131072")
                 MCA oob: parameter "oob_tcp_rcvbuf" (current value: "131072")
                 MCA oob: parameter "oob_tcp_connect_timeout" (current value: "10")
                          connect() timeout in seconds, before trying next interface
                 MCA oob: parameter "oob_tcp_listen_mode" (current value: "event")
                          Mode for HNP to accept incoming connections: event, listen_thread
                 MCA oob: parameter "oob_tcp_listen_thread_max_queue" (current value: "10")
                          High water mark for queued accepted socket list size
                 MCA oob: parameter "oob_tcp_listen_thread_max_time" (current value: "10")
                          Maximum amount of time (in milliseconds) to wait between processing accepted socket list
                 MCA oob: parameter "oob_tcp_accept_spin_count" (current value: "10")
                          Number of times to let accept return EWOULDBLOCK before updating accepted socket list
                 MCA oob: parameter "oob_tcp_priority" (current value: "0")
                 MCA ras: parameter "ras" (current value: <none>)
                 MCA ras: parameter "ras_dash_host_priority" (current value: "5")
                          Selection priority for the dash_host RAS component
                 MCA ras: parameter "ras_gridengine_debug" (current value: "0")
                          Enable debugging output for the gridengine ras component
                 MCA ras: parameter "ras_gridengine_priority" (current value: "100")
                          Priority of the gridengine ras component
                 MCA ras: parameter "ras_gridengine_verbose" (current value: "0")
                          Enable verbose output for the gridengine ras component
                 MCA ras: parameter "ras_gridengine_show_jobid" (current value: "0")
                          Show the JOB_ID of the Grid Engine job
                 MCA ras: parameter "ras_localhost_priority" (current value: "0")
                          Selection priority for the localhost RAS component
                 MCA ras: parameter "ras_slurm_priority" (current value: "75")
                          Priority of the slurm ras component
                 MCA rds: parameter "rds" (current value: <none>)
                 MCA rds: parameter "rds_hostfile_debug" (current value: "0")
                          Toggle debug output for hostfile RDS component
                 MCA rds: parameter "rds_hostfile_path" (current value: "/g/g12/mlleinin/src/ompi-v1.2b-112506-gcc/etc/openmpi-default-hostfile")
                          ORTE Host filename
                 MCA rds: parameter "rds_hostfile_priority" (current value: "0")
                 MCA rds: parameter "rds_proxy_priority" (current value: "0")
                 MCA rds: parameter "rds_resfile_debug" (current value: "0")
                          Toggle debug output for resfile RDS component
                 MCA rds: parameter "rds_resfile_name" (current value: <none>)
                          ORTE Resource filename
                 MCA rds: parameter "rds_resfile_priority" (current value: "0")
               MCA rmaps: parameter "rmaps_base_verbose" (current value: "0")
                          Verbosity level for the rmaps framework
               MCA rmaps: parameter "rmaps_base_schedule_policy" (current value: "slot")
                          Scheduling Policy for RMAPS. [slot | node]
               MCA rmaps: parameter "rmaps_base_pernode" (current value: "0")
                          Request one ppn if num procs not specified
               MCA rmaps: parameter "rmaps_base_schedule_local" (current value: "1")
                          If nonzero, allow scheduling MPI applications on the same node as mpirun (default). If zero, do not schedule any MPI applications on the same node as mpirun
               MCA rmaps: parameter "rmaps_base_no_oversubscribe" (current value: "0")
                          If nonzero, then do not allow oversubscription of nodes - mpirun will return an error if there aren't enough nodes to launch all processes without oversubscribing
               MCA rmaps: parameter "rmaps" (current value: <none>)
                          Default selection set of components for the rmaps framework (<none> means "use all components that can be found")
               MCA rmaps: parameter "rmaps_proxy_debug" (current value: "0")
               MCA rmaps: parameter "rmaps_proxy_priority" (current value: "0")
               MCA rmaps: parameter "rmaps_round_robin_debug" (current value: "1")
                          Toggle debug output for Round Robin RMAPS component
               MCA rmaps: parameter "rmaps_round_robin_priority" (current value: "1")
                          Selection priority for Round Robin RMAPS component
                MCA rmgr: parameter "rmgr" (current value: <none>)
                          Default selection set of components for the rmgr framework (<none> means "use all components that can be found")
                MCA rmgr: parameter "rmgr_proxy_priority" (current value: "0")
                MCA rmgr: parameter "rmgr_urm_priority" (current value: "0")
                 MCA rml: parameter "rml" (current value: <none>)
                          Default selection set of components for the rml framework (<none> means "use all components that can be found")
                 MCA rml: parameter "rml_base_verbose" (current value: "0")
                          Verbosity level for the rml framework (0 = no verbosity)
                 MCA rml: parameter "rml_oob_priority" (current value: "0")
                 MCA pls: parameter "pls" (current value: <none>)
                          Default selection set of components for the pls framework (<none> means "use all components that can be found")
                 MCA pls: parameter "pls_base_verbose" (current value: "0")
                          Verbosity level for the pls framework (0 = no verbosity)
                 MCA pls: parameter "pls_gridengine_debug" (current value: "0")
                          Enable debugging of gridengine pls component
                 MCA pls: parameter "pls_gridengine_verbose" (current value: "0")
                          Enable verbose output of the gridengine qrsh -inherit command
                 MCA pls: parameter "pls_gridengine_priority" (current value: "100")
                          Priority of the gridengine pls component
                 MCA pls: parameter "pls_gridengine_orted" (current value: "orted")
                          The command name that the gridengine pls component will invoke for the ORTE daemon
                 MCA pls: parameter "pls_proxy_priority" (current value: "0")
                 MCA pls: parameter "pls_rsh_debug" (current value: "0")
                          Whether or not to enable debugging output for the rsh pls component (0 or 1)
                 MCA pls: parameter "pls_rsh_num_concurrent" (current value: "128")
                          How many pls_rsh_agent instances to invoke concurrently (must be > 0)
                 MCA pls: parameter "pls_rsh_orted" (current value: "orted")
                          The command name that the rsh pls component will invoke for the ORTE daemon
                 MCA pls: parameter "pls_rsh_priority" (current value: "10")
                          Priority of the rsh pls component
                 MCA pls: parameter "pls_rsh_delay" (current value: "1")
                          Delay (in seconds) between invocations of the remote agent, but only used when the "debug" MCA parameter is true, or the top-level MCA debugging is enabled (otherwise this value is ignored)
                 MCA pls: parameter "pls_rsh_reap" (current value: "1")
                          If set to 1, wait for all the processes to complete before exiting. Otherwise, quit immediately -- without waiting for confirmation that all other processes in the job have completed.
                 MCA pls: parameter "pls_rsh_assume_same_shell" (current value: "1")
                          If set to 1, assume that the shell on the remote node is the same as the shell on the local node. Otherwise, probe for what the remote shell.
                 MCA pls: parameter "pls_rsh_agent" (current value: "ssh : rsh")
                          The command used to launch executables on remote nodes (typically either "ssh" or "rsh")
                 MCA pls: parameter "pls_slurm_debug" (current value: "0")
                          Enable debugging of slurm pls
                 MCA pls: parameter "pls_slurm_priority" (current value: "75")
                          Default selection priority
                 MCA pls: parameter "pls_slurm_orted" (current value: "orted")
                          Command to use to start proxy orted
                 MCA pls: parameter "pls_slurm_args" (current value: <none>)
                          Custom arguments to srun
                 MCA sds: parameter "sds" (current value: <none>)
                          Default selection set of components for the sds framework (<none> means "use all components that can be found")
                 MCA sds: parameter "sds_base_verbose" (current value: "0")
                          Verbosity level for the sds framework (0 = no verbosity)
                 MCA sds: parameter "sds_env_priority" (current value: "0")
                 MCA sds: parameter "sds_pipe_priority" (current value: "0")
                 MCA sds: parameter "sds_seed_priority" (current value: "0")
                 MCA sds: parameter "sds_singleton_priority" (current value: "0")
                 MCA sds: parameter "sds_slurm_priority" (current value: "0")