2007-07-20 05:34:02 +04:00
|
|
|
/*
|
2008-05-06 22:08:45 +04:00
|
|
|
* Copyright (c) 2004-2008 The Trustees of Indiana University and Indiana
|
2007-07-20 05:34:02 +04:00
|
|
|
* 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.
|
|
|
|
|
* 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:
|
|
|
|
|
*
|
|
|
|
|
* The OpenRTE Group Communications
|
|
|
|
|
*
|
|
|
|
|
* The OpenRTE Group Comm framework provides communication services that
|
|
|
|
|
* span entire jobs or collections of processes. It is not intended to be
|
|
|
|
|
* used for point-to-point communications (the RML does that), nor should
|
|
|
|
|
* it be viewed as a high-performance communication channel for large-scale
|
|
|
|
|
* data transfers.
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#ifndef MCA_GRPCOMM_H
|
|
|
|
|
#define MCA_GRPCOMM_H
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* includes
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
#include "orte_config.h"
|
2008-02-28 04:57:57 +03:00
|
|
|
#include "orte/constants.h"
|
|
|
|
|
#include "orte/types.h"
|
2007-07-20 05:34:02 +04:00
|
|
|
|
|
|
|
|
#include "opal/mca/mca.h"
|
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
|
|
|
#include "opal/class/opal_list.h"
|
2008-02-28 04:57:57 +03:00
|
|
|
#include "opal/dss/dss_types.h"
|
2008-04-10 02:10:53 +04:00
|
|
|
|
2007-07-20 05:34:02 +04:00
|
|
|
#include "orte/mca/rml/rml_types.h"
|
|
|
|
|
|
2008-02-28 04:57:57 +03:00
|
|
|
#include "orte/mca/grpcomm/grpcomm_types.h"
|
2007-07-20 05:34:02 +04:00
|
|
|
|
2008-02-28 04:57:57 +03:00
|
|
|
BEGIN_C_DECLS
|
2007-07-20 05:34:02 +04:00
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Component functions - all MUST be provided!
|
|
|
|
|
*/
|
|
|
|
|
|
2008-03-24 23:50:31 +03:00
|
|
|
/* initialize the selected module */
|
|
|
|
|
typedef int (*orte_grpcomm_base_module_init_fn_t)(void);
|
|
|
|
|
|
|
|
|
|
/* finalize the selected module */
|
|
|
|
|
typedef void (*orte_grpcomm_base_module_finalize_fn_t)(void);
|
|
|
|
|
|
2007-07-20 05:34:02 +04:00
|
|
|
/* Send a message to all members of a job - blocking */
|
|
|
|
|
typedef int (*orte_grpcomm_base_module_xcast_fn_t)(orte_jobid_t job,
|
2008-02-28 04:57:57 +03:00
|
|
|
opal_buffer_t *buffer,
|
2007-07-20 05:34:02 +04:00
|
|
|
orte_rml_tag_t tag);
|
|
|
|
|
|
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
|
|
|
/* allgather - gather data from all procs */
|
2008-02-28 04:57:57 +03:00
|
|
|
typedef int (*orte_grpcomm_base_module_allgather_fn_t)(opal_buffer_t *sbuf, opal_buffer_t *rbuf);
|
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
|
|
|
|
|
|
|
|
typedef int (*orte_grpcomm_base_module_allgather_list_fn_t)(opal_list_t *names,
|
2008-02-28 04:57:57 +03:00
|
|
|
opal_buffer_t *sbuf, opal_buffer_t *rbuf);
|
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
|
|
|
|
|
|
|
|
/* barrier function */
|
|
|
|
|
typedef int (*orte_grpcomm_base_module_barrier_fn_t)(void);
|
2007-07-20 05:34:02 +04:00
|
|
|
|
2009-05-11 18:11:44 +04:00
|
|
|
/* one-sided barrier function - process releases once its
|
|
|
|
|
* contribution is complete
|
|
|
|
|
*/
|
|
|
|
|
typedef int (*orte_grpcomm_base_module_onesided_barrier_fn_t)(void);
|
|
|
|
|
|
|
|
|
|
|
2008-02-28 04:57:57 +03:00
|
|
|
/** DATA EXCHANGE FUNCTIONS - SEE ompi/runtime/ompi_module_exchange.h FOR A DESCRIPTION
|
|
|
|
|
* OF HOW THIS ALL WORKS
|
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
/* send an attribute buffer */
|
2008-03-24 23:50:31 +03:00
|
|
|
typedef int (*orte_grpcomm_base_module_modex_set_proc_attr_fn_t)(const char* attr_name,
|
2008-04-10 02:10:53 +04:00
|
|
|
const void *buffer, size_t size);
|
2008-02-28 04:57:57 +03:00
|
|
|
|
|
|
|
|
/* get an attribute buffer */
|
2008-03-24 23:50:31 +03:00
|
|
|
typedef int (*orte_grpcomm_base_module_modex_get_proc_attr_fn_t)(const orte_process_name_t name,
|
2008-04-10 02:10:53 +04:00
|
|
|
const char* attr_name,
|
|
|
|
|
void **buffer, size_t *size);
|
2008-02-28 04:57:57 +03:00
|
|
|
|
|
|
|
|
/* perform a modex operation */
|
2008-03-24 23:50:31 +03:00
|
|
|
typedef int (*orte_grpcomm_base_module_modex_fn_t)(opal_list_t *procs);
|
2008-02-28 04:57:57 +03:00
|
|
|
|
|
|
|
|
/* purge the internal attr table */
|
2008-03-24 23:50:31 +03:00
|
|
|
typedef int (*orte_grpcomm_base_module_purge_proc_attrs_fn_t)(void);
|
2008-02-28 04:57:57 +03:00
|
|
|
|
|
|
|
|
|
2007-07-20 05:34:02 +04:00
|
|
|
/*
|
|
|
|
|
* Ver 2.0
|
|
|
|
|
*/
|
|
|
|
|
struct orte_grpcomm_base_module_2_0_0_t {
|
2008-03-24 23:50:31 +03:00
|
|
|
orte_grpcomm_base_module_init_fn_t init;
|
|
|
|
|
orte_grpcomm_base_module_finalize_fn_t finalize;
|
|
|
|
|
/* collective operations */
|
|
|
|
|
orte_grpcomm_base_module_xcast_fn_t xcast;
|
|
|
|
|
orte_grpcomm_base_module_allgather_fn_t allgather;
|
|
|
|
|
orte_grpcomm_base_module_allgather_list_fn_t allgather_list;
|
|
|
|
|
orte_grpcomm_base_module_barrier_fn_t barrier;
|
2009-05-11 18:11:44 +04:00
|
|
|
orte_grpcomm_base_module_onesided_barrier_fn_t onesided_barrier;
|
2008-03-24 23:50:31 +03:00
|
|
|
/* modex functions */
|
|
|
|
|
orte_grpcomm_base_module_modex_set_proc_attr_fn_t set_proc_attr;
|
|
|
|
|
orte_grpcomm_base_module_modex_get_proc_attr_fn_t get_proc_attr;
|
|
|
|
|
orte_grpcomm_base_module_modex_fn_t modex;
|
|
|
|
|
orte_grpcomm_base_module_purge_proc_attrs_fn_t purge_proc_attrs;
|
2007-07-20 05:34:02 +04:00
|
|
|
};
|
|
|
|
|
|
|
|
|
|
typedef struct orte_grpcomm_base_module_2_0_0_t orte_grpcomm_base_module_2_0_0_t;
|
|
|
|
|
typedef orte_grpcomm_base_module_2_0_0_t orte_grpcomm_base_module_t;
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* the standard component data structure
|
|
|
|
|
*/
|
|
|
|
|
struct orte_grpcomm_base_component_2_0_0_t {
|
2008-05-06 22:08:45 +04:00
|
|
|
mca_base_component_t base_version;
|
2008-07-29 02:40:57 +04:00
|
|
|
mca_base_component_data_t base_data;
|
2007-07-20 05:34:02 +04:00
|
|
|
};
|
|
|
|
|
typedef struct orte_grpcomm_base_component_2_0_0_t orte_grpcomm_base_component_2_0_0_t;
|
|
|
|
|
typedef orte_grpcomm_base_component_2_0_0_t orte_grpcomm_base_component_t;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
|
* Macro for use in components that are of type grpcomm v2.0.0
|
|
|
|
|
*/
|
|
|
|
|
#define ORTE_GRPCOMM_BASE_VERSION_2_0_0 \
|
2008-07-29 02:40:57 +04:00
|
|
|
/* grpcomm v2.0 is chained to MCA v2.0 */ \
|
|
|
|
|
MCA_BASE_VERSION_2_0_0, \
|
2007-07-20 05:34:02 +04:00
|
|
|
/* grpcomm v2.0 */ \
|
|
|
|
|
"grpcomm", 2, 0, 0
|
|
|
|
|
|
|
|
|
|
/* Global structure for accessing name server functions
|
|
|
|
|
*/
|
|
|
|
|
ORTE_DECLSPEC extern orte_grpcomm_base_module_t orte_grpcomm; /* holds selected module's function pointers */
|
|
|
|
|
|
2008-02-28 04:57:57 +03:00
|
|
|
END_C_DECLS
|
2007-07-20 05:34:02 +04:00
|
|
|
|
|
|
|
|
#endif
|