2004-08-14 05:56:05 +04:00
|
|
|
/*
|
2005-11-05 22:57:48 +03:00
|
|
|
* Copyright (c) 2004-2005 The Trustees of Indiana University and Indiana
|
|
|
|
|
* University Research and Technology
|
|
|
|
|
* Corporation. All rights reserved.
|
|
|
|
|
* Copyright (c) 2004-2005 The University of Tennessee and The University
|
|
|
|
|
* of Tennessee Research Foundation. All rights
|
|
|
|
|
* reserved.
|
2004-11-28 23:09:25 +03:00
|
|
|
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
|
|
|
|
|
* University of Stuttgart. All rights reserved.
|
2005-03-24 15:43:37 +03:00
|
|
|
* Copyright (c) 2004-2005 The Regents of the University of California.
|
|
|
|
|
* All rights reserved.
|
2007-02-09 23:17:37 +03:00
|
|
|
* Copyright (c) 2007 Los Alamos National Security, LLC. All rights
|
|
|
|
|
* reserved.
|
2007-01-21 17:02:06 +03:00
|
|
|
* Copyright (c) 2006-2007 Cisco Systems, Inc. All rights reserved.
|
2004-11-22 04:38:40 +03:00
|
|
|
* $COPYRIGHT$
|
|
|
|
|
*
|
|
|
|
|
* Additional copyrights may follow
|
|
|
|
|
*
|
2004-08-14 05:56:05 +04:00
|
|
|
* $HEADER$
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#ifndef OMPI_RUNTIME_PARAMS_H
|
|
|
|
|
#define OMPI_RUNTIME_PARAMS_H
|
2007-05-01 08:49:36 +04:00
|
|
|
|
|
|
|
|
BEGIN_C_DECLS
|
2004-08-14 05:56:05 +04:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Global variables
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether or not to check the parameters of top-level MPI API
|
|
|
|
|
* functions or not.
|
|
|
|
|
*
|
|
|
|
|
* This variable should never be checked directly; the macro
|
|
|
|
|
* MPI_PARAM_CHECK should be used instead. This allows multiple
|
|
|
|
|
* levels of MPI function parameter checking:
|
|
|
|
|
*
|
|
|
|
|
* #- Disable all parameter checking at configure/compile time
|
|
|
|
|
* #- Enable all parameter checking at configure/compile time
|
|
|
|
|
* #- Disable all parameter checking at run time
|
|
|
|
|
* #- Enable all parameter checking at run time
|
|
|
|
|
*
|
|
|
|
|
* Hence, the MPI_PARAM_CHECK macro will either be "0", "1", or
|
|
|
|
|
* "ompi_mpi_param_check".
|
|
|
|
|
*/
|
2004-10-22 20:06:05 +04:00
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_param_check;
|
2004-08-14 05:56:05 +04:00
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether or not to check for MPI handle leaks during MPI_FINALIZE.
|
|
|
|
|
* If enabled, each MPI handle type will display a summary of the
|
|
|
|
|
* handles that are still allocated during MPI_FINALIZE.
|
|
|
|
|
*
|
|
|
|
|
* This is good debugging for user applications to find out if they
|
|
|
|
|
* are inadvertantly orphaning MPI handles.
|
|
|
|
|
*/
|
2004-10-22 20:06:05 +04:00
|
|
|
OMPI_DECLSPEC extern bool ompi_debug_show_handle_leaks;
|
2004-08-14 05:56:05 +04:00
|
|
|
|
2007-08-02 01:33:25 +04:00
|
|
|
/**
|
|
|
|
|
* If > 0, show that many MPI_ALLOC_MEM leaks during MPI_FINALIZE. If
|
|
|
|
|
* enabled, memory that was returned via MPI_ALLOC_MEM but was never
|
|
|
|
|
* freed via MPI_FREE_MEM will be displayed during MPI_FINALIZE.
|
|
|
|
|
*
|
|
|
|
|
* This is good debugging for user applications to find out if they
|
|
|
|
|
* are inadvertantly orphaning MPI "special" memory.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern int ompi_debug_show_mpi_alloc_mem_leaks;
|
|
|
|
|
|
2004-08-14 05:56:05 +04:00
|
|
|
/**
|
|
|
|
|
* Whether or not to actually free MPI handles when their
|
|
|
|
|
* corresponding destructor is invoked. If enabled, Open MPI will not
|
|
|
|
|
* free handles, but will rather simply mark them as "freed". Any
|
|
|
|
|
* attempt to use them will result in an MPI exception.
|
|
|
|
|
*
|
|
|
|
|
* This is good debugging for user applications to find out if they
|
|
|
|
|
* are inadvertantly using MPI handles after they have been freed.
|
|
|
|
|
*/
|
2004-10-22 20:06:05 +04:00
|
|
|
OMPI_DECLSPEC extern bool ompi_debug_no_free_handles;
|
2004-08-14 05:56:05 +04:00
|
|
|
|
2005-02-21 21:56:30 +03:00
|
|
|
/**
|
|
|
|
|
* Whether or not to print MCA parameters on MPI_INIT
|
|
|
|
|
*
|
|
|
|
|
* This is good debugging for user applications to see exactly which
|
|
|
|
|
* MCA parameters are being used in the current program execution.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_show_mca_params;
|
2004-08-14 05:56:05 +04:00
|
|
|
|
2005-02-21 21:56:30 +03:00
|
|
|
/**
|
|
|
|
|
* Whether or not to print the MCA parameters to a file or to stdout
|
|
|
|
|
*
|
|
|
|
|
* If this argument is set then it is used when parameters are dumped
|
|
|
|
|
* when the mpi_show_mca_params is set.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern char * ompi_mpi_show_mca_params_file;
|
|
|
|
|
|
2005-08-16 20:17:52 +04:00
|
|
|
/**
|
|
|
|
|
* If this value is true, assume that this ORTE job is the only job
|
|
|
|
|
* running on the nodes that have been allocated to it, and bind
|
|
|
|
|
* processes to processors (starting with processor 0).
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_paffinity_alone;
|
|
|
|
|
|
2008-03-18 12:39:07 +03:00
|
|
|
/**
|
|
|
|
|
* If this value is true, we can check process binding to CPU
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool rmaps_rank_file_debug;
|
|
|
|
|
|
2007-05-01 08:49:36 +04:00
|
|
|
/**
|
|
|
|
|
* Whether we should keep the string hostnames of all the MPI
|
|
|
|
|
* process peers around or not (eats up a good bit of memory).
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_keep_peer_hostnames;
|
|
|
|
|
|
These changes were mostly captured in a prior RFC (except for #2 below) and are aimed specifically at improving startup performance and setting up the remaining modifications described in that RFC.
The commit has been tested for C/R and Cray operations, and on Odin (SLURM, rsh) and RoadRunner (TM). I tried to update all environments, but obviously could not test them. I know that Windows needs some work, and have highlighted what is know to be needed in the odls process component.
This represents a lot of work by Brian, Tim P, Josh, and myself, with much advice from Jeff and others. For posterity, I have appended a copy of the email describing the work that was done:
As we have repeatedly noted, the modex operation in MPI_Init is the single greatest consumer of time during startup. To-date, we have executed that operation as an ORTE stage gate that held the process until a startup message containing all required modex (and OOB contact info - see #3 below) info could be sent to it. Each process would send its data to the HNP's registry, which assembled and sent the message when all processes had reported in.
In addition, ORTE had taken responsibility for monitoring process status as it progressed through a series of "stage gates". The process reported its status at each gate, and ORTE would then send a "release" message once all procs had reported in.
The incoming changes revamp these procedures in three ways:
1. eliminating the ORTE stage gate system and cleanly delineating responsibility between the OMPI and ORTE layers for MPI init/finalize. The modex stage gate (STG1) has been replaced by a collective operation in the modex itself that performs an allgather on the required modex info. The allgather is implemented using the orte_grpcomm framework since the BTL's are not active at that point. At the moment, the grpcomm framework only has a "basic" component analogous to OMPI's "basic" coll framework - I would recommend that the MPI team create additional, more advanced components to improve performance of this step.
The other stage gates have been replaced by orte_grpcomm barrier functions. We tried to use MPI barriers instead (since the BTL's are active at that point), but - as we discussed on the telecon - these are not currently true barriers so the job would hang when we fell through while messages were still in process. Note that the grpcomm barrier doesn't actually resolve that problem, but Brian has pointed out that we are unlikely to ever see it violated. Again, you might want to spend a little time on an advanced barrier algorithm as the one in "basic" is very simplistic.
Summarizing this change: ORTE no longer tracks process state nor has direct responsibility for synchronizing jobs. This is now done via collective operations within the MPI layer, albeit using ORTE collective communication services. I -strongly- urge the MPI team to implement advanced collective algorithms to improve the performance of this critical procedure.
2. reducing the volume of data exchanged during modex. Data in the modex consisted of the process name, the name of the node where that process is located (expressed as a string), plus a string representation of all contact info. The nodename was required in order for the modex to determine if the process was local or not - in addition, some people like to have it to print pretty error messages when a connection failed.
The size of this data has been reduced in three ways:
(a) reducing the size of the process name itself. The process name consisted of two 32-bit fields for the jobid and vpid. This is far larger than any current system, or system likely to exist in the near future, can support. Accordingly, the default size of these fields has been reduced to 16-bits, which means you can have 32k procs in each of 32k jobs. Since the daemons must have a vpid, and we require one daemon/node, this also restricts the default configuration to 32k nodes.
To support any future "mega-clusters", a configuration option --enable-jumbo-apps has been added. This option increases the jobid and vpid field sizes to 32-bits. Someday, if necessary, someone can add yet another option to increase them to 64-bits, I suppose.
(b) replacing the string nodename with an integer nodeid. Since we have one daemon/node, the nodeid corresponds to the local daemon's vpid. This replaces an often lengthy string with only 2 (or at most 4) bytes, a substantial reduction.
(c) when the mca param requesting that nodenames be sent to support pretty error messages, a second mca param is now used to request FQDN - otherwise, the domain name is stripped (by default) from the message to save space. If someone wants to combine those into a single param somehow (perhaps with an argument?), they are welcome to do so - I didn't want to alter what people are already using.
While these may seem like small savings, they actually amount to a significant impact when aggregated across the entire modex operation. Since every proc must receive the modex data regardless of the collective used to send it, just reducing the size of the process name removes nearly 400MBytes of communication from a 32k proc job (admittedly, much of this comm may occur in parallel). So it does add up pretty quickly.
3. routing RML messages to reduce connections. The default messaging system remains point-to-point - i.e., each proc opens a socket to every proc it communicates with and sends its messages directly. A new option uses the orteds as routers - i.e., each proc only opens a single socket to its local orted. All messages are sent from the proc to the orted, which forwards the message to the orted on the node where the intended recipient proc is located - that orted then forwards the message to its local proc (the recipient). This greatly reduces the connection storm we have encountered during startup.
It also has the benefit of removing the sharing of every proc's OOB contact with every other proc. The orted routing tables are populated during launch since every orted gets a map of where every proc is being placed. Each proc, therefore, only needs to know the contact info for its local daemon, which is passed in via the environment when the proc is fork/exec'd by the daemon. This alone removes ~50 bytes/process of communication that was in the current STG1 startup message - so for our 32k proc job, this saves us roughly 32k*50 = 1.6MBytes sent to 32k procs = 51GBytes of messaging.
Note that you can use the new routing method by specifying -mca routed tree - if you so desire. This mode will become the default at some point in the future.
There are a few minor additional changes in the commit that I'll just note in passing:
* propagation of command line mca params to the orteds - fixes ticket #1073. See note there for details.
* requiring of "finalize" prior to "exit" for MPI procs - fixes ticket #1144. See note there for details.
* cleanup of some stale header files
This commit was SVN r16364.
2007-10-05 23:48:23 +04:00
|
|
|
/**
|
|
|
|
|
* Whether or not to use the FQDN for the peer hostnames. This
|
|
|
|
|
* can eat up a good bit of memory as well as a lot of communication
|
|
|
|
|
* during startup - it can be reduced by just using the hostname
|
|
|
|
|
* instead of the FQDN
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_keep_fqdn_hostnames;
|
|
|
|
|
|
2007-05-01 08:49:36 +04:00
|
|
|
/**
|
|
|
|
|
* Whether an MPI_ABORT should print out a stack trace or not.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_abort_print_stack;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether MPI_ABORT should print out an identifying message
|
|
|
|
|
* (e.g., hostname and PID) and loop waiting for a debugger to
|
|
|
|
|
* attach. The value of the integer is how many seconds to wait:
|
|
|
|
|
*
|
|
|
|
|
* 0 = do not print the message and do not loop
|
|
|
|
|
* negative value = print the message and loop forever
|
|
|
|
|
* positive value = print the message and delay for that many seconds
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern int ompi_mpi_abort_delay;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether to use the "leave pinned" protocol or not.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_leave_pinned;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether to use the "leave pinned pipeline" protocol or not.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_mpi_leave_pinned_pipeline;
|
|
|
|
|
|
2007-08-04 04:41:26 +04:00
|
|
|
/**
|
|
|
|
|
* Whether sparse MPI group storage formats are supported or not.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_have_sparse_group_storage;
|
|
|
|
|
|
|
|
|
|
/**
|
|
|
|
|
* Whether sparse MPI group storage formats should be used or not.
|
|
|
|
|
*/
|
|
|
|
|
OMPI_DECLSPEC extern bool ompi_use_sparse_group_storage;
|
|
|
|
|
|
2007-05-01 08:49:36 +04:00
|
|
|
/**
|
|
|
|
|
* Register MCA parameters used by the MPI layer.
|
|
|
|
|
*
|
|
|
|
|
* @returns OMPI_SUCCESS
|
|
|
|
|
*
|
|
|
|
|
* Registers several MCA parameters and initializes corresponding
|
|
|
|
|
* global variables to the values obtained from the MCA system.
|
|
|
|
|
*/
|
2005-02-10 22:08:35 +03:00
|
|
|
OMPI_DECLSPEC int ompi_mpi_register_params(void);
|
2005-02-21 21:56:30 +03:00
|
|
|
|
2006-03-31 04:31:15 +04:00
|
|
|
|
2007-05-01 08:49:36 +04:00
|
|
|
/**
|
|
|
|
|
* Display all MCA parameters used
|
|
|
|
|
*
|
|
|
|
|
* @returns OMPI_SUCCESS
|
|
|
|
|
*
|
|
|
|
|
* Displays in key = value format
|
|
|
|
|
*/
|
|
|
|
|
int ompi_show_all_mca_params(int32_t, int, char *);
|
|
|
|
|
|
|
|
|
|
END_C_DECLS
|
2004-08-14 05:56:05 +04:00
|
|
|
|
|
|
|
|
#endif /* OMPI_RUNTIME_PARAMS_H */
|