54b2cf747e
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 The modex stage gate (STG1) has been replaced by a collective operation in the modex itself that performs an allgather on the required modex info.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 The other stage gates have been replaced by orte_grpcomm barrier functions.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.
204 строки
6.2 KiB
C
204 строки
6.2 KiB
C
/*
|
|
* Copyright (c) 2004-2007 The Trustees of Indiana University and Indiana
|
|
* University Research and Technology
|
|
* Corporation. All rights reserved.
|
|
* Copyright (c) 2004-2006 The University of Tennessee and The University
|
|
* of Tennessee Research Foundation. All rights
|
|
* reserved.
|
|
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
|
|
* University of Stuttgart. All rights reserved.
|
|
* Copyright (c) 2004-2005 The Regents of the University of California.
|
|
* All rights reserved.
|
|
* $COPYRIGHT$
|
|
*
|
|
* Additional copyrights may follow
|
|
*
|
|
* $HEADER$
|
|
*/
|
|
/** @file:
|
|
*
|
|
* Contains the internal functions and typedefs for the use of the oob
|
|
*/
|
|
|
|
#ifndef MCA_OOB_H_
|
|
#define MCA_OOB_H_
|
|
|
|
#include "orte_config.h"
|
|
|
|
#ifdef HAVE_UNISTD_H
|
|
#include <unistd.h>
|
|
#endif
|
|
|
|
#include "opal/types.h"
|
|
#include "opal/mca/mca.h"
|
|
|
|
#include "orte/mca/ns/ns_types.h"
|
|
|
|
#include "orte/mca/rml/rml.h"
|
|
#include "orte/mca/oob/oob_types.h"
|
|
|
|
#include "opal/mca/crs/crs.h"
|
|
#include "opal/mca/crs/base/base.h"
|
|
|
|
|
|
BEGIN_C_DECLS
|
|
|
|
|
|
struct mca_oob_1_0_0_t;
|
|
typedef struct mca_oob_1_0_0_t mca_oob_1_0_0_t;
|
|
typedef struct mca_oob_1_0_0_t mca_oob_t;
|
|
|
|
|
|
typedef mca_oob_t* (*mca_oob_base_component_init_fn_t)(int *priority);
|
|
|
|
struct mca_oob_base_component_1_0_0_t {
|
|
mca_base_component_t oob_base;
|
|
mca_base_component_data_1_0_0_t oob_data;
|
|
mca_oob_base_component_init_fn_t oob_init;
|
|
};
|
|
typedef struct mca_oob_base_component_1_0_0_t mca_oob_base_component_1_0_0_t;
|
|
typedef mca_oob_base_component_1_0_0_t mca_oob_base_component_t;
|
|
|
|
|
|
|
|
typedef char* (*mca_oob_base_module_get_addr_fn_t)(void);
|
|
|
|
typedef int (*mca_oob_base_module_set_addr_fn_t)(const orte_process_name_t* peer,
|
|
const char* uri);
|
|
|
|
|
|
typedef int (*mca_oob_base_module_get_new_name_fn_t)(orte_process_name_t*);
|
|
|
|
/**
|
|
* Implementation of mca_oob_ping().
|
|
*
|
|
* @param peer (IN) Opaque name of peer process.
|
|
* @param tv (IN) Timeout to wait in connection response.
|
|
* @return OMPI error code (<0) or ORTE_SUCCESS
|
|
*/
|
|
typedef int (*mca_oob_base_module_ping_fn_t)(const orte_process_name_t*,
|
|
const char* uri,
|
|
const struct timeval* tv);
|
|
|
|
|
|
/**
|
|
* Send an oob message
|
|
*
|
|
* Send an oob message. All oob sends are non-blocking, and cbfunc
|
|
* will be called when the message has been sent. When cbfunc is
|
|
* called, message has been injected into the network but no guarantee
|
|
* is made about whether the target has received the message.
|
|
*
|
|
* @param[in] target Destination process name
|
|
* @param[in] origin Origin process for the message, for the purposes
|
|
* of message matching. This can be different from
|
|
* the process calling send().
|
|
* @param[in] msg Array of iovecs describing user buffers and lengths.
|
|
* @param[in] count Number of elements in iovec array.
|
|
* @param[in] tag User defined tag for matching send/recv.
|
|
* @param[in] flags Currently unused.
|
|
* @param[in] cbfunc Callback function on send completion.
|
|
* @param[in] cbdata User data that is passed to callback function.
|
|
*
|
|
* @return OMPI error code (<0) on error number of bytes actually sent.
|
|
*/
|
|
typedef int (*mca_oob_base_module_send_nb_fn_t)(
|
|
orte_process_name_t* target,
|
|
orte_process_name_t* origin,
|
|
struct iovec* msg,
|
|
int count,
|
|
int tag,
|
|
int flags,
|
|
orte_rml_callback_fn_t cbfunc,
|
|
void* cbdata);
|
|
|
|
/**
|
|
* Implementation of mca_oob_recv_nb().
|
|
*
|
|
* @param peer (IN) Opaque name of peer process or ORTE_NAME_WILDCARD for wildcard receive.
|
|
* @param msg (IN) Array of iovecs describing user buffers and lengths.
|
|
* @param count (IN) Number of elements in iovec array.
|
|
* @param tag (IN) User defined tag for matching send/recv.
|
|
* @param flags (IN) May be MCA_OOB_PEEK to return up to size bytes of msg w/out removing it from the queue,
|
|
* @param cbfunc (IN) Callback function on recv completion.
|
|
* @param cbdata (IN) User data that is passed to callback function.
|
|
* @return OMPI error code (<0) on error or number of bytes actually received.
|
|
*/
|
|
|
|
typedef int (*mca_oob_base_module_recv_nb_fn_t)(
|
|
orte_process_name_t* peer,
|
|
struct iovec* msg,
|
|
int count,
|
|
int tag,
|
|
int flags,
|
|
orte_rml_callback_fn_t cbfunc,
|
|
void* cbdata);
|
|
|
|
/**
|
|
* Implementation of mca_oob_recv_cancel().
|
|
*
|
|
* @param peer (IN) Opaque name of peer process or ORTE_NAME_WILDCARD for wildcard receive.
|
|
* @param tag (IN) User defined tag for matching send/recv.
|
|
* @return OMPI error code (<0) on error or number of bytes actually received.
|
|
*/
|
|
|
|
typedef int (*mca_oob_base_module_recv_cancel_fn_t)(orte_process_name_t* peer, int tag);
|
|
|
|
/**
|
|
* Hook function called by mca_oob_base_register to allow
|
|
* the oob component a chance to register contact information
|
|
*/
|
|
typedef int (*mca_oob_base_module_init_fn_t)(void);
|
|
|
|
/**
|
|
* Cleanup during finalize.
|
|
*/
|
|
typedef int (*mca_oob_base_module_fini_fn_t)(void);
|
|
|
|
/* ft event */
|
|
typedef int (*mca_oob_base_module_ft_event_fn_t)( int state );
|
|
|
|
|
|
/**
|
|
* OOB Module
|
|
*/
|
|
struct mca_oob_1_0_0_t {
|
|
mca_oob_base_module_init_fn_t oob_init;
|
|
mca_oob_base_module_fini_fn_t oob_fini;
|
|
|
|
mca_oob_base_module_get_addr_fn_t oob_get_addr;
|
|
mca_oob_base_module_set_addr_fn_t oob_set_addr;
|
|
|
|
mca_oob_base_module_get_new_name_fn_t oob_get_new_name;
|
|
mca_oob_base_module_ping_fn_t oob_ping;
|
|
|
|
mca_oob_base_module_send_nb_fn_t oob_send_nb;
|
|
|
|
mca_oob_base_module_recv_nb_fn_t oob_recv_nb;
|
|
mca_oob_base_module_recv_cancel_fn_t oob_recv_cancel;
|
|
|
|
mca_oob_base_module_ft_event_fn_t oob_ft_event;
|
|
|
|
orte_rml_exception_callback_t oob_exception_callback;
|
|
};
|
|
|
|
/**
|
|
* Macro for use in components that are of type oob v1.0.0
|
|
*/
|
|
#define MCA_OOB_BASE_VERSION_1_0_0 \
|
|
/* oob v1.0 is chained to MCA v1.0 */ \
|
|
MCA_BASE_VERSION_1_0_0, \
|
|
/* oob v1.0 */ \
|
|
"oob", 1, 0, 0
|
|
|
|
/*
|
|
* BWB - FIX ME - This is the first module on the list. This is here
|
|
* temporarily to make things work
|
|
*/
|
|
|
|
ORTE_DECLSPEC extern mca_oob_t mca_oob;
|
|
|
|
END_C_DECLS
|
|
|
|
#endif
|