ports/openmpi
ports/openmpi
1
1
Форкнуть 0
Код Релизы Активность
00842ad87c
openmpi/ompi/mca/coll/coll.h

179 строки
6.7 KiB
C
Исходник Обычный вид История

First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/*
First cut at copyrights: IU, UTK, and some OSU. LANL and HLRS still pending. This commit was SVN r3655.
2004-11-22 04:38:40 +03:00
* Copyright (c) 2004-2005 The Trustees of Indiana University.
* All rights reserved.
* Copyright (c) 2004-2005 The Trustees of the University of Tennessee.
* All rights reserved.
Add HLRS copyright This commit was SVN r3665.
2004-11-28 23:09:25 +03:00
* Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
* University of Stuttgart. All rights reserved.
Add UC copyright This commit was SVN r5009.
2005-03-24 15:43:37 +03:00
* Copyright (c) 2004-2005 The Regents of the University of California.
* All rights reserved.
First cut at copyrights: IU, UTK, and some OSU. LANL and HLRS still pending. This commit was SVN r3655.
2004-11-22 04:38:40 +03:00
* $COPYRIGHT$
*
* Additional copyrights may follow
*
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
* $HEADER$
*/
#ifndef MCA_COLL_H
#define MCA_COLL_H
First pass of lam -> ompi conversion This commit was SVN r1191.
2004-06-07 19:33:53 +04:00
#include "ompi_config.h"
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
#include "mpi.h"
#include "mca/mca.h"
Updating header file names after the great directory reorganization. This commit was SVN r877.
2004-03-17 21:45:16 +03:00
#include "mca/base/base.h"
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
Slightly change the coll comm query API -- pass in a **data to allow the component to cache data that is necessary to do the query (e.g., if you need to try to allocate resources in order to determine selectability). This data is given back during unquery() and is cached on comm->c_coll_selected_data if the module is selected (before calling comm_init). This commit was SVN r4251.
2005-01-30 21:42:37 +03:00
/*
* Forward declaration
*/
struct mca_coll_base_comm_t;
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/*
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
* Coll component function typedefs
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
*/
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef int (*mca_coll_base_component_init_query_fn_t)
Change the orientation of the thread level determination in ompi_mpi_init(): we no longer *query* the components to see what they can support -- instead, we *tell* the components what they need to support (if they can't support it, they should disqualify themselves from selection). This involved a bunch of things: - pass the configure-time-decided OMPI_ENABLE_PROGRESS_THREADS and OMPI_ENABLE_MPI_THREADS constants to each framework selection function - eliminate mca_base_init_select_components() - eliminate all "allow_multi_user_threads" and "have_hidden_threads" kinds of arguments and aggregation from various framework selection and component/module functions - correctly determine the max MPI thread level that we can support and limit the final MPI thread level as appropriate during ompi_mpi_init() As a side effect, while editing nearly every MPI component I also did the following: - found 2 places (gm and ib ptls) where ompi_set_using_threads() was erroneously being called (I think this is left over kruft from misunderstandings by the original authors). I removed these. This commit was SVN r5055.
2005-03-27 17:05:23 +04:00
(bool enable_progress_threads, bool enable_mpi_threads);
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef const struct mca_coll_base_module_1_0_0_t *
(*mca_coll_base_component_comm_query_1_0_0_fn_t)
Slightly change the coll comm query API -- pass in a **data to allow the component to cache data that is necessary to do the query (e.g., if you need to try to allocate resources in order to determine selectability). This data is given back during unquery() and is cached on comm->c_coll_selected_data if the module is selected (before calling comm_init). This commit was SVN r4251.
2005-01-30 21:42:37 +03:00
(struct ompi_communicator_t *comm, int *priority,
struct mca_coll_base_comm_t **data);
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef int (*mca_coll_base_component_comm_unquery_fn_t)
Slightly change the coll comm query API -- pass in a **data to allow the component to cache data that is necessary to do the query (e.g., if you need to try to allocate resources in order to determine selectability). This data is given back during unquery() and is cached on comm->c_coll_selected_data if the module is selected (before calling comm_init). This commit was SVN r4251.
2005-01-30 21:42:37 +03:00
(struct ompi_communicator_t *comm,
struct mca_coll_base_comm_t *data);
Lots of changes in this commit, mostly having to do with the first real commit of the collectives. MPI_SCAN and MPI_EXSCAN are still not implemented, but lots of other things are in the critical path and holding up other people, so it's ok to commit without them: - better checks for sizes in configure, and add defaults for fortran sizes if we don't have a fortran compiler - fix some logic that was accidentally broken for size checks for the file type offset_t - add some C equivalent types for fortran's complex and double complex (for use in internal reduction/op functions) - additionals and slight reorganization of ompi_mpi_init() ompi_mpi_finalize() - fully implement all top-level MPI collective calls, including all param checking for both intra- and inter-communicators (woof) - change the communicator_t type for stuff that we need in coll, and update all references throughout the code base to match - all kinds of updates to the coll framework base - next cut of the basic coll module -- has all intracommunicator collectives implemented except scan and exscan (see note above). All intercommunicator functions return ERR_NOT_IMPLEMENTED. - MPI_Op is a fixed implementation -- not component-ized yet. So there are generic C loops for all implementations. This commit was SVN r1491.
2004-06-29 04:02:25 +04:00
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
/*
* Coll module function typedefs
*/
typedef const struct mca_coll_base_module_1_0_0_t *
Lots of changes in this commit, mostly having to do with the first real commit of the collectives. MPI_SCAN and MPI_EXSCAN are still not implemented, but lots of other things are in the critical path and holding up other people, so it's ok to commit without them: - better checks for sizes in configure, and add defaults for fortran sizes if we don't have a fortran compiler - fix some logic that was accidentally broken for size checks for the file type offset_t - add some C equivalent types for fortran's complex and double complex (for use in internal reduction/op functions) - additionals and slight reorganization of ompi_mpi_init() ompi_mpi_finalize() - fully implement all top-level MPI collective calls, including all param checking for both intra- and inter-communicators (woof) - change the communicator_t type for stuff that we need in coll, and update all references throughout the code base to match - all kinds of updates to the coll framework base - next cut of the basic coll module -- has all intracommunicator collectives implemented except scan and exscan (see note above). All intercommunicator functions return ERR_NOT_IMPLEMENTED. - MPI_Op is a fixed implementation -- not component-ized yet. So there are generic C loops for all implementations. This commit was SVN r1491.
2004-06-29 04:02:25 +04:00
(*mca_coll_base_module_init_1_0_0_fn_t)(struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_finalize_fn_t)
(struct ompi_communicator_t *comm);
First cut of a lot of MCA infrastruture. Still little base code yet, but the build structure is getting there. This commit was SVN r269.
2004-01-12 00:31:52 +03:00
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef int (*mca_coll_base_module_allgather_fn_t)
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
void *rbuf, int rcount, struct ompi_datatype_t *rdtype,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_allgatherv_fn_t)
I hate waiting on the airports. SO I start doing something usefull ... I remove a lot of inter-dependence, I use the struct_t type. BEWARE not all the function are ready. This commit was SVN r3524.
2004-11-05 10:52:30 +03:00
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
void * rbuf, int *rcounts, int *disps, struct ompi_datatype_t *rdtype,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_allreduce_fn_t)
(void *sbuf, void *rbuf, int count, struct ompi_datatype_t *dtype,
struct ompi_op_t *op, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_alltoall_fn_t)
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
void* rbuf, int rcount, struct ompi_datatype_t *rdtype,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_alltoallv_fn_t)
(void *sbuf, int *scounts, int *sdisps, struct ompi_datatype_t *sdtype,
void *rbuf, int *rcounts, int *rdisps, struct ompi_datatype_t *rdtype,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_alltoallw_fn_t)
(void *sbuf, int *scounts, int *sdisps, struct ompi_datatype_t **sdtypes,
void *rbuf, int *rcounts, int *rdisps, struct ompi_datatype_t **rdtypes,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_barrier_fn_t)
(struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_bcast_fn_t)
(void *buff, int count, struct ompi_datatype_t *datatype, int root,
struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_exscan_fn_t)
(void *sbuf, void *rbuf, int count, struct ompi_datatype_t *dtype,
struct ompi_op_t *op, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_gather_fn_t)
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
void *rbuf, int rcount, struct ompi_datatype_t *rdtype,
int root, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_gatherv_fn_t)
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
void *rbuf, int *rcounts, int *disps, struct ompi_datatype_t *rdtype,
int root, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_reduce_fn_t)
(void *sbuf, void* rbuf, int count, struct ompi_datatype_t *dtype,
struct ompi_op_t *op, int root, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_reduce_scatter_fn_t)
(void *sbuf, void *rbuf, int *rcounts, struct ompi_datatype_t *dtype,
struct ompi_op_t *op, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_scan_fn_t)
(void *sbuf, void *rbuf, int count, struct ompi_datatype_t *dtype,
struct ompi_op_t *op, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_scatter_fn_t)
(void *sbuf, int scount, struct ompi_datatype_t *sdtype,
void *rbuf, int rcount, struct ompi_datatype_t *rdtype,
int root, struct ompi_communicator_t *comm);
typedef int (*mca_coll_base_module_scatterv_fn_t)
(void *sbuf, int *scounts, int *disps, struct ompi_datatype_t *sdtype,
void* rbuf, int rcount, struct ompi_datatype_t *rdtype,
int root, struct ompi_communicator_t *comm);
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/*
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
* Structure for coll v1.0.0 components
First cut of a lot of MCA infrastruture. Still little base code yet, but the build structure is getting there. This commit was SVN r269.
2004-01-12 00:31:52 +03:00
* Chained to MCA v1.0.0
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
*/
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
struct mca_coll_base_component_1_0_0_t {
mca_base_component_t collm_version;
mca_base_component_data_1_0_0_t collm_data;
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/* Initialization / querying functions */
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
mca_coll_base_component_init_query_fn_t collm_init_query;
mca_coll_base_component_comm_query_1_0_0_fn_t collm_comm_query;
mca_coll_base_component_comm_unquery_fn_t collm_comm_unquery;
- Various updates - Make names doxygen-friendly This commit was SVN r211.
2004-01-10 20:27:28 +03:00
};
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef struct mca_coll_base_component_1_0_0_t mca_coll_base_component_1_0_0_t;
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/*
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
* This struct is hung on the communicator by the winning coll component
- Renamed the main struct to "module", and rename the "actions" struct to the main struct - Rename the prefixes This commit was SVN r168.
2004-01-09 20:01:57 +03:00
* after the negotiation. It has pointers for all the collective
* functions, as well as a "finalize" function for when the
* communicator is freed.
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
*/
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
struct mca_coll_base_module_1_0_0_t {
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/* Per-communicator initialization and finalization functions */
Lots of changes in this commit, mostly having to do with the first real commit of the collectives. MPI_SCAN and MPI_EXSCAN are still not implemented, but lots of other things are in the critical path and holding up other people, so it's ok to commit without them: - better checks for sizes in configure, and add defaults for fortran sizes if we don't have a fortran compiler - fix some logic that was accidentally broken for size checks for the file type offset_t - add some C equivalent types for fortran's complex and double complex (for use in internal reduction/op functions) - additionals and slight reorganization of ompi_mpi_init() ompi_mpi_finalize() - fully implement all top-level MPI collective calls, including all param checking for both intra- and inter-communicators (woof) - change the communicator_t type for stuff that we need in coll, and update all references throughout the code base to match - all kinds of updates to the coll framework base - next cut of the basic coll module -- has all intracommunicator collectives implemented except scan and exscan (see note above). All intercommunicator functions return ERR_NOT_IMPLEMENTED. - MPI_Op is a fixed implementation -- not component-ized yet. So there are generic C loops for all implementations. This commit was SVN r1491.
2004-06-29 04:02:25 +04:00
mca_coll_base_module_init_1_0_0_fn_t coll_module_init;
mca_coll_base_module_finalize_fn_t coll_module_finalize;
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
/* Collective function pointers */
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
mca_coll_base_module_allgather_fn_t coll_allgather;
mca_coll_base_module_allgatherv_fn_t coll_allgatherv;
mca_coll_base_module_allreduce_fn_t coll_allreduce;
mca_coll_base_module_alltoall_fn_t coll_alltoall;
mca_coll_base_module_alltoallv_fn_t coll_alltoallv;
mca_coll_base_module_alltoallw_fn_t coll_alltoallw;
mca_coll_base_module_barrier_fn_t coll_barrier;
mca_coll_base_module_bcast_fn_t coll_bcast;
mca_coll_base_module_exscan_fn_t coll_exscan;
mca_coll_base_module_gather_fn_t coll_gather;
mca_coll_base_module_gatherv_fn_t coll_gatherv;
mca_coll_base_module_reduce_fn_t coll_reduce;
mca_coll_base_module_reduce_scatter_fn_t coll_reduce_scatter;
mca_coll_base_module_scan_fn_t coll_scan;
mca_coll_base_module_scatter_fn_t coll_scatter;
mca_coll_base_module_scatterv_fn_t coll_scatterv;
- Various updates - Make names doxygen-friendly This commit was SVN r211.
2004-01-10 20:27:28 +03:00
};
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
typedef struct mca_coll_base_module_1_0_0_t mca_coll_base_module_1_0_0_t;
First cut of a lot of MCA infrastruture. Still little base code yet, but the build structure is getting there. This commit was SVN r269.
2004-01-12 00:31:52 +03:00
/*
- massive change for module<-->component name fixes throughout the code base. - many (most) mca type names have "component" or "module" in them, as relevant, just to further distinguish the difference between component data/actions and module data/actions. All developers are encouraged to perpetuate this convention when you create types that are specific to a framework, component, or module - did very little to entire framework (just the basics to make it compile) because it's just about to be almost entirely replaced - ditto for io / romio - did not work on elan or ib components; have to commit and then convert those on a different machine with the right libraries and headers - renamed a bunch of *_module.c files to *_component.c and *module*c to *component*c (a few still remain, e.g., ptl/ib, ptl/elan, etc.) - modified autogen/configure/build process to match new filenames (e.g., output static-components.h instead of static-modules.h) - removed DOS-style cr/lf stuff in ns/ns.h - added newline to end of file src/util/numtostr.h - removed some redundant error checking in the top-level topo functions - added a few {} here and there where people "forgot" to put them in for 1 line blocks ;-) - removed a bunch of MPI_* types from mca header files (replaced with corresponding ompi_* types) - all the ptl components had version numbers in their structs; removed - converted a few more elements in the MCA base to use the OBJ interface -- removed some old manual reference counting kruft This commit was SVN r1830.
2004-08-02 04:24:22 +04:00
* Macro for use in components that are of type coll v1.0.0
First cut of a lot of MCA infrastruture. Still little base code yet, but the build structure is getting there. This commit was SVN r269.
2004-01-12 00:31:52 +03:00
*/
#define MCA_COLL_BASE_VERSION_1_0_0 \
/* coll v1.0 is chained to MCA v1.0 */ \
MCA_BASE_VERSION_1_0_0, \
/* coll v1.0 */ \
"coll", 1, 0, 0
First take at the collective interface This commit was SVN r145.
2004-01-09 08:21:01 +03:00
#endif /* MCA_COLL_H */
Копировать постоянную ссылку