2009-07-01 01:41:28 +04:00
|
|
|
/*
|
A number of C/R enhancements per RFC below:
http://www.open-mpi.org/community/lists/devel/2010/07/8240.php
Documentation:
http://osl.iu.edu/research/ft/
Major Changes:
--------------
* Added C/R-enabled Debugging support.
Enabled with the --enable-crdebug flag. See the following website for more information:
http://osl.iu.edu/research/ft/crdebug/
* Added Stable Storage (SStore) framework for checkpoint storage
* 'central' component does a direct to central storage save
* 'stage' component stages checkpoints to central storage while the application continues execution.
* 'stage' supports offline compression of checkpoints before moving (sstore_stage_compress)
* 'stage' supports local caching of checkpoints to improve automatic recovery (sstore_stage_caching)
* Added Compression (compress) framework to support
* Add two new ErrMgr recovery policies
* {{{crmig}}} C/R Process Migration
* {{{autor}}} C/R Automatic Recovery
* Added the {{{ompi-migrate}}} command line tool to support the {{{crmig}}} ErrMgr component
* Added CR MPI Ext functions (enable them with {{{--enable-mpi-ext=cr}}} configure option)
* {{{OMPI_CR_Checkpoint}}} (Fixes trac:2342)
* {{{OMPI_CR_Restart}}}
* {{{OMPI_CR_Migrate}}} (may need some more work for mapping rules)
* {{{OMPI_CR_INC_register_callback}}} (Fixes trac:2192)
* {{{OMPI_CR_Quiesce_start}}}
* {{{OMPI_CR_Quiesce_checkpoint}}}
* {{{OMPI_CR_Quiesce_end}}}
* {{{OMPI_CR_self_register_checkpoint_callback}}}
* {{{OMPI_CR_self_register_restart_callback}}}
* {{{OMPI_CR_self_register_continue_callback}}}
* The ErrMgr predicted_fault() interface has been changed to take an opal_list_t of ErrMgr defined types. This will allow us to better support a wider range of fault prediction services in the future.
* Add a progress meter to:
* FileM rsh (filem_rsh_process_meter)
* SnapC full (snapc_full_progress_meter)
* SStore stage (sstore_stage_progress_meter)
* Added 2 new command line options to ompi-restart
* --showme : Display the full command line that would have been exec'ed.
* --mpirun_opts : Command line options to pass directly to mpirun. (Fixes trac:2413)
* Deprecated some MCA params:
* crs_base_snapshot_dir deprecated, use sstore_stage_local_snapshot_dir
* snapc_base_global_snapshot_dir deprecated, use sstore_base_global_snapshot_dir
* snapc_base_global_shared deprecated, use sstore_stage_global_is_shared
* snapc_base_store_in_place deprecated, replaced with different components of SStore
* snapc_base_global_snapshot_ref deprecated, use sstore_base_global_snapshot_ref
* snapc_base_establish_global_snapshot_dir deprecated, never well supported
* snapc_full_skip_filem deprecated, use sstore_stage_skip_filem
Minor Changes:
--------------
* Fixes trac:1924 : {{{ompi-restart}}} now recognizes path prefixed checkpoint handles and does the right thing.
* Fixes trac:2097 : {{{ompi-info}}} should now report all available CRS components
* Fixes trac:2161 : Manual checkpoint movement. A user can 'mv' a checkpoint directory from the original location to another and still restart from it.
* Fixes trac:2208 : Honor various TMPDIR varaibles instead of forcing {{{/tmp}}}
* Move {{{ompi_cr_continue_like_restart}}} to {{{orte_cr_continue_like_restart}}} to be more flexible in where this should be set.
* opal_crs_base_metadata_write* functions have been moved to SStore to support a wider range of metadata handling functionality.
* Cleanup the CRS framework and components to work with the SStore framework.
* Cleanup the SnapC framework and components to work with the SStore framework (cleans up these code paths considerably).
* Add 'quiesce' hook to CRCP for a future enhancement.
* We now require a BLCR version that supports {{{cr_request_file()}}} or {{{cr_request_checkpoint()}}} in order to make the code more maintainable. Note that {{{cr_request_file}}} has been deprecated since 0.7.0, so we prefer to use {{{cr_request_checkpoint()}}}.
* Add optional application level INC callbacks (registered through the CR MPI Ext interface).
* Increase the {{{opal_cr_thread_sleep_wait}}} parameter to 1000 microseconds to make the C/R thread less aggressive.
* {{{opal-restart}}} now looks for cache directories before falling back on stable storage when asked.
* {{{opal-restart}}} also support local decompression before restarting
* {{{orte-checkpoint}}} now uses the SStore framework to work with the metadata
* {{{orte-restart}}} now uses the SStore framework to work with the metadata
* Remove the {{{orte-restart}}} preload option. This was removed since the user only needs to select the 'stage' component in order to support this functionality.
* Since the '-am' parameter is saved in the metadata, {{{ompi-restart}}} no longer hard codes {{{-am ft-enable-cr}}}.
* Fix {{{hnp}}} ErrMgr so that if a previous component in the stack has 'fixed' the problem, then it should be skipped.
* Make sure to decrement the number of 'num_local_procs' in the orted when one goes away.
* odls now checks the SStore framework to see if it needs to load any checkpoint files before launching (to support 'stage'). This separates the SStore logic from the --preload-[binary|files] options.
* Add unique IDs to the named pipes established between the orted and the app in SnapC. This is to better support migration and automatic recovery activities.
* Improve the checks for 'already checkpointing' error path.
* A a recovery output timer, to show how long it takes to restart a job
* Do a better job of cleaning up the old session directory on restart.
* Add a local module to the autor and crmig ErrMgr components. These small modules prevent the 'orted' component from attempting a local recovery (Which does not work for MPI apps at the moment)
* Add a fix for bounding the checkpointable region between MPI_Init and MPI_Finalize.
This commit was SVN r23587.
The following Trac tickets were found above:
Ticket 1924 --> https://svn.open-mpi.org/trac/ompi/ticket/1924
Ticket 2097 --> https://svn.open-mpi.org/trac/ompi/ticket/2097
Ticket 2161 --> https://svn.open-mpi.org/trac/ompi/ticket/2161
Ticket 2192 --> https://svn.open-mpi.org/trac/ompi/ticket/2192
Ticket 2208 --> https://svn.open-mpi.org/trac/ompi/ticket/2208
Ticket 2342 --> https://svn.open-mpi.org/trac/ompi/ticket/2342
Ticket 2413 --> https://svn.open-mpi.org/trac/ompi/ticket/2413
2010-08-11 00:51:11 +04:00
|
|
|
* Copyright (c) 2004-2010 The Trustees of Indiana University and Indiana
|
2009-07-01 01:41:28 +04:00
|
|
|
* 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 (c) 2007-2009 Cisco Systems, Inc. All rights reserved.
|
|
|
|
* Copyright (c) 2009 Oak Ridge National Labs. All rights reserved.
|
|
|
|
* $COPYRIGHT$
|
|
|
|
*
|
|
|
|
* Additional copyrights may follow
|
|
|
|
*
|
|
|
|
* $HEADER$
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include "ompi_config.h"
|
|
|
|
#include "mpi.h"
|
|
|
|
|
2009-07-01 17:19:06 +04:00
|
|
|
#include <string.h>
|
2009-07-01 01:41:28 +04:00
|
|
|
#include <ctype.h>
|
2009-07-16 22:27:33 +04:00
|
|
|
#ifdef HAVE_UNISTD_H
|
2009-07-01 01:41:28 +04:00
|
|
|
#include <unistd.h>
|
|
|
|
#endif
|
2009-07-16 22:27:33 +04:00
|
|
|
#ifdef HAVE_SYS_PARAM_H
|
2009-07-01 01:41:28 +04:00
|
|
|
#include <sys/param.h>
|
|
|
|
#endif
|
|
|
|
#ifdef HAVE_NETDB_H
|
|
|
|
#include <netdb.h>
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#include MCA_timer_IMPLEMENTATION_HEADER
|
|
|
|
#include "opal/mca/installdirs/installdirs.h"
|
|
|
|
#include "opal/class/opal_value_array.h"
|
|
|
|
#include "opal/class/opal_pointer_array.h"
|
|
|
|
#include "opal/util/printf.h"
|
|
|
|
#include "opal/memoryhooks/memory.h"
|
|
|
|
#include "opal/mca/base/mca_base_param.h"
|
|
|
|
|
|
|
|
#include "orte/util/show_help.h"
|
|
|
|
|
|
|
|
#include "ompi/tools/ompi_info/ompi_info.h"
|
|
|
|
#include "ompi/include/mpi_portable_platform.h"
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Public variables
|
|
|
|
*/
|
|
|
|
|
|
|
|
const char *ompi_info_component_all = "all";
|
|
|
|
const char *ompi_info_param_all = "all";
|
|
|
|
|
|
|
|
const char *ompi_info_path_prefix = "prefix";
|
|
|
|
const char *ompi_info_path_bindir = "bindir";
|
|
|
|
const char *ompi_info_path_libdir = "libdir";
|
|
|
|
const char *ompi_info_path_incdir = "incdir";
|
|
|
|
const char *ompi_info_path_mandir = "mandir";
|
|
|
|
const char *ompi_info_path_pkglibdir = "pkglibdir";
|
|
|
|
const char *ompi_info_path_sysconfdir = "sysconfdir";
|
|
|
|
const char *ompi_info_path_exec_prefix = "exec_prefix";
|
|
|
|
const char *ompi_info_path_sbindir = "sbindir";
|
|
|
|
const char *ompi_info_path_libexecdir = "libexecdir";
|
|
|
|
const char *ompi_info_path_datarootdir = "datarootdir";
|
|
|
|
const char *ompi_info_path_datadir = "datadir";
|
|
|
|
const char *ompi_info_path_sharedstatedir = "sharedstatedir";
|
|
|
|
const char *ompi_info_path_localstatedir = "localstatedir";
|
|
|
|
const char *ompi_info_path_infodir = "infodir";
|
|
|
|
const char *ompi_info_path_pkgdatadir = "pkgdatadir";
|
|
|
|
const char *ompi_info_path_pkgincludedir = "pkgincludedir";
|
|
|
|
|
|
|
|
/*
|
|
|
|
* External variables
|
|
|
|
*
|
|
|
|
* This exists in mca/base/mca_base_param.c. It's not extern'ed
|
|
|
|
* in mca_base_param.h so that no one else will use it.
|
|
|
|
*/
|
|
|
|
|
|
|
|
extern opal_value_array_t mca_base_params;
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_do_params(bool want_all_in, bool want_internal)
|
|
|
|
{
|
|
|
|
int count;
|
|
|
|
char *type, *component, *str;
|
|
|
|
bool found;
|
|
|
|
int i;
|
|
|
|
bool want_all = false;
|
|
|
|
opal_list_t *info;
|
|
|
|
|
|
|
|
ompi_info_open_components();
|
|
|
|
|
|
|
|
if (want_all_in) {
|
|
|
|
want_all = true;
|
|
|
|
} else {
|
|
|
|
/* See if the special param "all" was givin to --param; that
|
|
|
|
* superceeds any individual type
|
|
|
|
*/
|
|
|
|
count = opal_cmd_line_get_ninsts(ompi_info_cmd_line, "param");
|
|
|
|
for (i = 0; i < count; ++i) {
|
|
|
|
type = opal_cmd_line_get_param(ompi_info_cmd_line, "param", (int)i, 0);
|
|
|
|
if (0 == strcmp(ompi_info_type_all, type)) {
|
|
|
|
want_all = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Get a dump of all the MCA params */
|
|
|
|
mca_base_param_dump(&info, want_internal);
|
|
|
|
|
|
|
|
/* Show the params */
|
|
|
|
|
|
|
|
if (want_all) {
|
|
|
|
for (i = 0; i < mca_types.size; ++i) {
|
2009-07-01 17:26:07 +04:00
|
|
|
if (NULL == (type = (char *)opal_pointer_array_get_item(&mca_types, i))) {
|
2009-07-01 01:41:28 +04:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
ompi_info_show_mca_params(info, type, ompi_info_component_all, want_internal);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
for (i = 0; i < count; ++i) {
|
|
|
|
type = opal_cmd_line_get_param(ompi_info_cmd_line, "param", (int)i, 0);
|
|
|
|
component = opal_cmd_line_get_param(ompi_info_cmd_line, "param", (int)i, 1);
|
|
|
|
|
|
|
|
for (found = false, i = 0; i < mca_types.size; ++i) {
|
2009-07-01 17:26:07 +04:00
|
|
|
if (NULL == (str = (char *)opal_pointer_array_get_item(&mca_types, i))) {
|
2009-07-01 01:41:28 +04:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
if (0 == strcmp(str, type)) {
|
|
|
|
found = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!found) {
|
|
|
|
char *usage = opal_cmd_line_get_usage_msg(ompi_info_cmd_line);
|
2009-09-29 17:56:14 +04:00
|
|
|
orte_show_help("help-ompi_info.txt", "not-found", true, type);
|
2009-07-01 01:41:28 +04:00
|
|
|
free(usage);
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
|
|
|
|
ompi_info_show_mca_params(info, type, component, want_internal);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Release all the MCA param memory */
|
|
|
|
mca_base_param_dump_release(info);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_show_mca_params(opal_list_t *info,
|
|
|
|
const char *type, const char *component,
|
|
|
|
bool want_internal)
|
|
|
|
{
|
|
|
|
opal_list_item_t *i;
|
|
|
|
mca_base_param_info_t *p;
|
|
|
|
char *value_string, *empty = "";
|
|
|
|
char *message, *content, *tmp;
|
|
|
|
int value_int, j;
|
|
|
|
mca_base_param_source_t source;
|
|
|
|
char *src_file;
|
|
|
|
|
|
|
|
for (i = opal_list_get_first(info); i != opal_list_get_last(info);
|
|
|
|
i = opal_list_get_next(i)) {
|
|
|
|
p = (mca_base_param_info_t*) i;
|
|
|
|
|
|
|
|
if (NULL != p->mbpp_type_name && 0 == strcmp(type, p->mbpp_type_name)) {
|
|
|
|
if (0 == strcmp(component, ompi_info_component_all) ||
|
|
|
|
NULL == p->mbpp_component_name ||
|
|
|
|
(NULL != p->mbpp_component_name &&
|
|
|
|
0 == strcmp(component, p->mbpp_component_name))) {
|
|
|
|
|
|
|
|
/* Find the source of the value */
|
|
|
|
if (OPAL_SUCCESS !=
|
|
|
|
mca_base_param_lookup_source(p->mbpp_index, &source, &src_file)) {
|
|
|
|
continue;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Make a char *for the default value. Invoke a
|
|
|
|
* lookup because it may transform the char *("~/" ->
|
|
|
|
* "<home dir>/") or get the value from the
|
|
|
|
* environment, a file, etc.
|
|
|
|
*/
|
|
|
|
if (MCA_BASE_PARAM_TYPE_STRING == p->mbpp_type) {
|
|
|
|
mca_base_param_lookup_string(p->mbpp_index,
|
|
|
|
&value_string);
|
|
|
|
|
|
|
|
/* Can't let the char *be NULL because we
|
|
|
|
* assign it to a std::string, below
|
|
|
|
*/
|
|
|
|
if (NULL == value_string) {
|
|
|
|
value_string = strdup(empty);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
mca_base_param_lookup_int(p->mbpp_index, &value_int);
|
|
|
|
asprintf(&value_string, "%d", value_int);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Build up the strings to ompi_info_output. */
|
|
|
|
|
|
|
|
if (ompi_info_pretty) {
|
|
|
|
asprintf(&message, "MCA %s", p->mbpp_type_name);
|
|
|
|
|
|
|
|
/* Put in the real, full name (which may be
|
|
|
|
* different than the categorization).
|
|
|
|
*/
|
|
|
|
asprintf(&content, "%s \"%s\" (%s: <%s>, data source: ",
|
|
|
|
p->mbpp_read_only ? "information" : "parameter",
|
|
|
|
p->mbpp_full_name,
|
|
|
|
p->mbpp_read_only ? "value" : "current value",
|
|
|
|
(0 == strlen(value_string)) ? "none" : value_string);
|
|
|
|
|
|
|
|
/* Indicate where the param was set from */
|
|
|
|
switch(source) {
|
|
|
|
case MCA_BASE_PARAM_SOURCE_DEFAULT:
|
|
|
|
asprintf(&tmp, "%sdefault value", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_ENV:
|
|
|
|
asprintf(&tmp, "%senvironment or cmdline", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_FILE:
|
|
|
|
asprintf(&tmp, "%sfile [%s]", content, src_file);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_OVERRIDE:
|
|
|
|
asprintf(&tmp, "%sAPI override", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is this parameter deprecated? */
|
|
|
|
if (p->mbpp_deprecated) {
|
|
|
|
asprintf(&tmp, "%s, deprecated", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Does this parameter have any synonyms? */
|
|
|
|
if (p->mbpp_synonyms_len > 0) {
|
|
|
|
asprintf(&tmp, "%s, synonyms: ", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
for (j = 0; j < p->mbpp_synonyms_len; ++j) {
|
|
|
|
if (j > 0) {
|
|
|
|
asprintf(&tmp, "%s, %s", content, p->mbpp_synonyms[j]->mbpp_full_name);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
} else {
|
|
|
|
asprintf(&tmp, "%s%s", content, p->mbpp_synonyms[j]->mbpp_full_name);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is this parameter a synonym of something else? */
|
|
|
|
else if (NULL != p->mbpp_synonym_parent) {
|
|
|
|
asprintf(&tmp, "%s, synonym of: %s", content, p->mbpp_synonym_parent->mbpp_full_name);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
}
|
|
|
|
asprintf(&tmp, "%s)", content);
|
|
|
|
free(content);
|
|
|
|
content = tmp;
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
free(content);
|
|
|
|
|
|
|
|
/* If we have a help message, ompi_info_output it */
|
|
|
|
if (NULL != p->mbpp_help_msg) {
|
|
|
|
ompi_info_out("", "", p->mbpp_help_msg);
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
/* build the message*/
|
|
|
|
asprintf(&tmp, "mca:%s:%s:param:%s:", p->mbpp_type_name,
|
|
|
|
(NULL == p->mbpp_component_name) ? "base" : p->mbpp_component_name,
|
|
|
|
p->mbpp_full_name);
|
|
|
|
|
|
|
|
/* Output the value */
|
|
|
|
asprintf(&message, "%svalue", tmp);
|
|
|
|
ompi_info_out(message, message, value_string);
|
|
|
|
free(message);
|
|
|
|
|
|
|
|
/* Indicate where the param was set from */
|
|
|
|
|
|
|
|
asprintf(&message, "%sdata_source", tmp);
|
|
|
|
switch(source) {
|
|
|
|
case MCA_BASE_PARAM_SOURCE_DEFAULT:
|
|
|
|
content = strdup("default value");
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_ENV:
|
|
|
|
content = strdup("environment-cmdline");
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_FILE:
|
|
|
|
asprintf(&content, "file: %s", src_file);
|
|
|
|
break;
|
|
|
|
case MCA_BASE_PARAM_SOURCE_OVERRIDE:
|
|
|
|
content = strdup("API override");
|
|
|
|
break;
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
free(content);
|
|
|
|
|
|
|
|
/* Output whether it's read only or writable */
|
|
|
|
|
|
|
|
asprintf(&message, "%sstatus", tmp);
|
|
|
|
content = p->mbpp_read_only ? "read-only" : "writable";
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
|
|
|
|
/* If it has a help message, ompi_info_output that */
|
|
|
|
|
|
|
|
if (NULL != p->mbpp_help_msg) {
|
|
|
|
asprintf(&message, "%shelp", tmp);
|
|
|
|
content = p->mbpp_help_msg;
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is this parameter deprecated? */
|
|
|
|
asprintf(&message, "%sdeprecated", tmp);
|
|
|
|
content = p->mbpp_deprecated ? "yes" : "no";
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
|
|
|
|
/* Does this parameter have any synonyms? */
|
|
|
|
if (p->mbpp_synonyms_len > 0) {
|
|
|
|
for (j = 0; j < p->mbpp_synonyms_len; ++j) {
|
|
|
|
asprintf(&message, "%ssynonym:name", tmp);
|
|
|
|
content = p->mbpp_synonyms[j]->mbpp_full_name;
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Is this parameter a synonym of something else? */
|
|
|
|
else if (NULL != p->mbpp_synonym_parent) {
|
|
|
|
asprintf(&message, "%ssynonym_of:name", tmp);
|
|
|
|
content = p->mbpp_synonym_parent->mbpp_full_name;
|
|
|
|
ompi_info_out(message, message, content);
|
|
|
|
free(message);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/* If we allocated the string, then free it */
|
|
|
|
|
|
|
|
if (NULL != value_string) {
|
|
|
|
free(value_string);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_do_path(bool want_all, opal_cmd_line_t *cmd_line)
|
|
|
|
{
|
|
|
|
int i, count;
|
|
|
|
char *scope;
|
|
|
|
|
|
|
|
/* Check bozo case */
|
|
|
|
count = opal_cmd_line_get_ninsts(cmd_line, "path");
|
|
|
|
for (i = 0; i < count; ++i) {
|
|
|
|
scope = opal_cmd_line_get_param(cmd_line, "path", i, 0);
|
|
|
|
if (0 == strcmp("all", scope)) {
|
|
|
|
want_all = true;
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (want_all) {
|
|
|
|
ompi_info_show_path(ompi_info_path_prefix, opal_install_dirs.prefix);
|
|
|
|
ompi_info_show_path(ompi_info_path_exec_prefix, opal_install_dirs.exec_prefix);
|
|
|
|
ompi_info_show_path(ompi_info_path_bindir, opal_install_dirs.bindir);
|
|
|
|
ompi_info_show_path(ompi_info_path_sbindir, opal_install_dirs.sbindir);
|
|
|
|
ompi_info_show_path(ompi_info_path_libdir, opal_install_dirs.libdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_incdir, opal_install_dirs.includedir);
|
|
|
|
ompi_info_show_path(ompi_info_path_mandir, opal_install_dirs.mandir);
|
|
|
|
ompi_info_show_path(ompi_info_path_pkglibdir, opal_install_dirs.pkglibdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_libexecdir, opal_install_dirs.libexecdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_datarootdir, opal_install_dirs.datarootdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_datadir, opal_install_dirs.datadir);
|
|
|
|
ompi_info_show_path(ompi_info_path_sysconfdir, opal_install_dirs.sysconfdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_sharedstatedir, opal_install_dirs.sharedstatedir);
|
|
|
|
ompi_info_show_path(ompi_info_path_localstatedir, opal_install_dirs.localstatedir);
|
|
|
|
ompi_info_show_path(ompi_info_path_infodir, opal_install_dirs.infodir);
|
|
|
|
ompi_info_show_path(ompi_info_path_pkgdatadir, opal_install_dirs.pkgdatadir);
|
|
|
|
ompi_info_show_path(ompi_info_path_pkglibdir, opal_install_dirs.pkglibdir);
|
|
|
|
ompi_info_show_path(ompi_info_path_pkgincludedir, opal_install_dirs.pkgincludedir);
|
|
|
|
} else {
|
|
|
|
count = opal_cmd_line_get_ninsts(cmd_line, "path");
|
|
|
|
for (i = 0; i < count; ++i) {
|
|
|
|
scope = opal_cmd_line_get_param(cmd_line, "path", i, 0);
|
|
|
|
|
|
|
|
if (0 == strcmp(ompi_info_path_prefix, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_prefix, opal_install_dirs.prefix);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_bindir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_bindir, opal_install_dirs.bindir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_libdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_libdir, opal_install_dirs.libdir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_incdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_incdir, opal_install_dirs.includedir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_mandir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_mandir, opal_install_dirs.mandir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_pkglibdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_pkglibdir, opal_install_dirs.pkglibdir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_sysconfdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_sysconfdir, opal_install_dirs.sysconfdir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_exec_prefix, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_exec_prefix, opal_install_dirs.exec_prefix);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_sbindir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_sbindir, opal_install_dirs.sbindir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_libexecdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_libexecdir, opal_install_dirs.libexecdir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_datarootdir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_datarootdir, opal_install_dirs.datarootdir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_datadir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_datadir, opal_install_dirs.datadir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_sharedstatedir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_sharedstatedir, opal_install_dirs.sharedstatedir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_localstatedir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_localstatedir, opal_install_dirs.localstatedir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_infodir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_infodir, opal_install_dirs.infodir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_pkgdatadir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_pkgdatadir, opal_install_dirs.pkgdatadir);
|
|
|
|
} else if (0 == strcmp(ompi_info_path_pkgincludedir, scope)) {
|
|
|
|
ompi_info_show_path(ompi_info_path_pkgincludedir, opal_install_dirs.pkgincludedir);
|
|
|
|
} else {
|
|
|
|
char *usage = opal_cmd_line_get_usage_msg(cmd_line);
|
|
|
|
orte_show_help("help-ompi_info.txt", "usage", true, usage);
|
|
|
|
free(usage);
|
|
|
|
exit(1);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_show_path(const char *type, const char *value)
|
|
|
|
{
|
|
|
|
char *pretty, *path;
|
|
|
|
|
|
|
|
pretty = strdup(type);
|
|
|
|
pretty[0] = toupper(pretty[0]);
|
|
|
|
|
|
|
|
asprintf(&path, "path:%s", type);
|
|
|
|
ompi_info_out(pretty, path, value);
|
|
|
|
free(pretty);
|
|
|
|
free(path);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_do_arch()
|
|
|
|
{
|
|
|
|
ompi_info_out("Configured architecture", "config:arch", OPAL_ARCH);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
void ompi_info_do_hostname()
|
|
|
|
{
|
2010-11-13 02:22:11 +03:00
|
|
|
ompi_info_out("Configure host", "config:host", OPAL_CONFIGURE_HOST);
|
2009-07-01 01:41:28 +04:00
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
/*
|
|
|
|
* do_config
|
|
|
|
* Accepts:
|
|
|
|
* - want_all: boolean flag; TRUE -> display all options
|
|
|
|
* FALSE -> display selected options
|
|
|
|
*
|
|
|
|
* This function displays all the options with which the current
|
|
|
|
* installation of ompi was configured. There are many options here
|
|
|
|
* that are carried forward from OMPI-7 and are not mca parameters
|
|
|
|
* in OMPI-10. I have to dig through the invalid options and replace
|
|
|
|
* them with OMPI-10 options.
|
|
|
|
*/
|
|
|
|
void ompi_info_do_config(bool want_all)
|
|
|
|
{
|
|
|
|
char *cxx;
|
|
|
|
char *f77;
|
|
|
|
char *f90;
|
|
|
|
char *f90_size;
|
|
|
|
char *heterogeneous;
|
|
|
|
char *memprofile;
|
|
|
|
char *memdebug;
|
|
|
|
char *debug;
|
2010-05-18 03:08:56 +04:00
|
|
|
char *log_event;
|
2009-07-01 01:41:28 +04:00
|
|
|
char *mpi_interface_warning;
|
|
|
|
char *cprofiling;
|
|
|
|
char *cxxprofiling;
|
|
|
|
char *f77profiling;
|
|
|
|
char *f90profiling;
|
|
|
|
char *cxxexceptions;
|
|
|
|
char *threads;
|
|
|
|
char *want_libltdl;
|
|
|
|
char *mpirun_prefix_by_default;
|
|
|
|
char *sparse_groups;
|
|
|
|
char *have_mpi_io;
|
|
|
|
char *wtime_support;
|
|
|
|
char *symbol_visibility;
|
|
|
|
char *ft_support;
|
A number of C/R enhancements per RFC below:
http://www.open-mpi.org/community/lists/devel/2010/07/8240.php
Documentation:
http://osl.iu.edu/research/ft/
Major Changes:
--------------
* Added C/R-enabled Debugging support.
Enabled with the --enable-crdebug flag. See the following website for more information:
http://osl.iu.edu/research/ft/crdebug/
* Added Stable Storage (SStore) framework for checkpoint storage
* 'central' component does a direct to central storage save
* 'stage' component stages checkpoints to central storage while the application continues execution.
* 'stage' supports offline compression of checkpoints before moving (sstore_stage_compress)
* 'stage' supports local caching of checkpoints to improve automatic recovery (sstore_stage_caching)
* Added Compression (compress) framework to support
* Add two new ErrMgr recovery policies
* {{{crmig}}} C/R Process Migration
* {{{autor}}} C/R Automatic Recovery
* Added the {{{ompi-migrate}}} command line tool to support the {{{crmig}}} ErrMgr component
* Added CR MPI Ext functions (enable them with {{{--enable-mpi-ext=cr}}} configure option)
* {{{OMPI_CR_Checkpoint}}} (Fixes trac:2342)
* {{{OMPI_CR_Restart}}}
* {{{OMPI_CR_Migrate}}} (may need some more work for mapping rules)
* {{{OMPI_CR_INC_register_callback}}} (Fixes trac:2192)
* {{{OMPI_CR_Quiesce_start}}}
* {{{OMPI_CR_Quiesce_checkpoint}}}
* {{{OMPI_CR_Quiesce_end}}}
* {{{OMPI_CR_self_register_checkpoint_callback}}}
* {{{OMPI_CR_self_register_restart_callback}}}
* {{{OMPI_CR_self_register_continue_callback}}}
* The ErrMgr predicted_fault() interface has been changed to take an opal_list_t of ErrMgr defined types. This will allow us to better support a wider range of fault prediction services in the future.
* Add a progress meter to:
* FileM rsh (filem_rsh_process_meter)
* SnapC full (snapc_full_progress_meter)
* SStore stage (sstore_stage_progress_meter)
* Added 2 new command line options to ompi-restart
* --showme : Display the full command line that would have been exec'ed.
* --mpirun_opts : Command line options to pass directly to mpirun. (Fixes trac:2413)
* Deprecated some MCA params:
* crs_base_snapshot_dir deprecated, use sstore_stage_local_snapshot_dir
* snapc_base_global_snapshot_dir deprecated, use sstore_base_global_snapshot_dir
* snapc_base_global_shared deprecated, use sstore_stage_global_is_shared
* snapc_base_store_in_place deprecated, replaced with different components of SStore
* snapc_base_global_snapshot_ref deprecated, use sstore_base_global_snapshot_ref
* snapc_base_establish_global_snapshot_dir deprecated, never well supported
* snapc_full_skip_filem deprecated, use sstore_stage_skip_filem
Minor Changes:
--------------
* Fixes trac:1924 : {{{ompi-restart}}} now recognizes path prefixed checkpoint handles and does the right thing.
* Fixes trac:2097 : {{{ompi-info}}} should now report all available CRS components
* Fixes trac:2161 : Manual checkpoint movement. A user can 'mv' a checkpoint directory from the original location to another and still restart from it.
* Fixes trac:2208 : Honor various TMPDIR varaibles instead of forcing {{{/tmp}}}
* Move {{{ompi_cr_continue_like_restart}}} to {{{orte_cr_continue_like_restart}}} to be more flexible in where this should be set.
* opal_crs_base_metadata_write* functions have been moved to SStore to support a wider range of metadata handling functionality.
* Cleanup the CRS framework and components to work with the SStore framework.
* Cleanup the SnapC framework and components to work with the SStore framework (cleans up these code paths considerably).
* Add 'quiesce' hook to CRCP for a future enhancement.
* We now require a BLCR version that supports {{{cr_request_file()}}} or {{{cr_request_checkpoint()}}} in order to make the code more maintainable. Note that {{{cr_request_file}}} has been deprecated since 0.7.0, so we prefer to use {{{cr_request_checkpoint()}}}.
* Add optional application level INC callbacks (registered through the CR MPI Ext interface).
* Increase the {{{opal_cr_thread_sleep_wait}}} parameter to 1000 microseconds to make the C/R thread less aggressive.
* {{{opal-restart}}} now looks for cache directories before falling back on stable storage when asked.
* {{{opal-restart}}} also support local decompression before restarting
* {{{orte-checkpoint}}} now uses the SStore framework to work with the metadata
* {{{orte-restart}}} now uses the SStore framework to work with the metadata
* Remove the {{{orte-restart}}} preload option. This was removed since the user only needs to select the 'stage' component in order to support this functionality.
* Since the '-am' parameter is saved in the metadata, {{{ompi-restart}}} no longer hard codes {{{-am ft-enable-cr}}}.
* Fix {{{hnp}}} ErrMgr so that if a previous component in the stack has 'fixed' the problem, then it should be skipped.
* Make sure to decrement the number of 'num_local_procs' in the orted when one goes away.
* odls now checks the SStore framework to see if it needs to load any checkpoint files before launching (to support 'stage'). This separates the SStore logic from the --preload-[binary|files] options.
* Add unique IDs to the named pipes established between the orted and the app in SnapC. This is to better support migration and automatic recovery activities.
* Improve the checks for 'already checkpointing' error path.
* A a recovery output timer, to show how long it takes to restart a job
* Do a better job of cleaning up the old session directory on restart.
* Add a local module to the autor and crmig ErrMgr components. These small modules prevent the 'orted' component from attempting a local recovery (Which does not work for MPI apps at the moment)
* Add a fix for bounding the checkpointable region between MPI_Init and MPI_Finalize.
This commit was SVN r23587.
The following Trac tickets were found above:
Ticket 1924 --> https://svn.open-mpi.org/trac/ompi/ticket/1924
Ticket 2097 --> https://svn.open-mpi.org/trac/ompi/ticket/2097
Ticket 2161 --> https://svn.open-mpi.org/trac/ompi/ticket/2161
Ticket 2192 --> https://svn.open-mpi.org/trac/ompi/ticket/2192
Ticket 2208 --> https://svn.open-mpi.org/trac/ompi/ticket/2208
Ticket 2342 --> https://svn.open-mpi.org/trac/ompi/ticket/2342
Ticket 2413 --> https://svn.open-mpi.org/trac/ompi/ticket/2413
2010-08-11 00:51:11 +04:00
|
|
|
char *crdebug_support;
|
2011-09-13 04:39:14 +04:00
|
|
|
char *hwloc_support;
|
2009-07-01 01:41:28 +04:00
|
|
|
/* Do a little preprocessor trickery here to figure ompi_info_out the
|
|
|
|
* tri-state of MPI_PARAM_CHECK (which will be either 0, 1, or
|
|
|
|
* ompi_mpi_param_check). The preprocessor will only allow
|
|
|
|
* comparisons against constants, so you'll get a warning if you
|
|
|
|
* check MPI_PARAM_CHECK against 0 or 1, but its real value is the
|
|
|
|
* char *ompi_mpi_param_check. So define ompi_mpi_param_check to
|
|
|
|
* be a constant, and then all the preprocessor comparisons work ompi_info_out
|
|
|
|
* ok. Note that we chose the preprocessor comparison rompi_info_oute because
|
|
|
|
* it is not sufficient to simply set the variable
|
|
|
|
* ompi_mpi_param_check to a non-0/non-1 value. This is because the
|
|
|
|
* compiler will generate a warning that that C variable is unused
|
|
|
|
* when MPI_PARAM_CHECK is hard-coded to 0 or 1.
|
|
|
|
*/
|
|
|
|
char *paramcheck;
|
|
|
|
#define ompi_mpi_param_check 999
|
|
|
|
#if 0 == MPI_PARAM_CHECK
|
|
|
|
paramcheck = "never";
|
|
|
|
#elif 1 == MPI_PARAM_CHECK
|
|
|
|
paramcheck = "always";
|
|
|
|
#else
|
|
|
|
paramcheck = "runtime";
|
|
|
|
#endif
|
|
|
|
|
|
|
|
/* setup the strings that don't require allocations*/
|
|
|
|
cxx = OMPI_WANT_CXX_BINDINGS ? "yes" : "no";
|
|
|
|
f90 = OMPI_WANT_F90_BINDINGS ? "yes" : "no";
|
|
|
|
f90_size = OMPI_F90_BUILD_SIZE;
|
|
|
|
heterogeneous = OPAL_ENABLE_HETEROGENEOUS_SUPPORT ? "yes" : "no";
|
|
|
|
memprofile = OPAL_ENABLE_MEM_PROFILE ? "yes" : "no";
|
|
|
|
memdebug = OPAL_ENABLE_MEM_DEBUG ? "yes" : "no";
|
|
|
|
debug = OPAL_ENABLE_DEBUG ? "yes" : "no";
|
2010-05-18 03:08:56 +04:00
|
|
|
log_event = ORTE_WANT_NOTIFIER_LOG_EVENT ? "yes" : "no";
|
2009-07-01 01:41:28 +04:00
|
|
|
mpi_interface_warning = OMPI_WANT_MPI_INTERFACE_WARNING ? "yes" : "no";
|
|
|
|
cprofiling = OMPI_ENABLE_MPI_PROFILING ? "yes" : "no";
|
2009-08-04 15:54:01 +04:00
|
|
|
cxxprofiling = (OMPI_WANT_CXX_BINDINGS && OMPI_ENABLE_MPI_PROFILING) ? "yes" : "no";
|
|
|
|
cxxexceptions = (OMPI_WANT_CXX_BINDINGS && OMPI_HAVE_CXX_EXCEPTION_SUPPORT) ? "yes" : "no";
|
2009-07-01 01:41:28 +04:00
|
|
|
f77profiling = (OMPI_ENABLE_MPI_PROFILING && OMPI_WANT_F77_BINDINGS) ? "yes" : "no";
|
|
|
|
f90profiling = (OMPI_ENABLE_MPI_PROFILING && OMPI_WANT_F90_BINDINGS) ? "yes" : "no";
|
|
|
|
want_libltdl = OPAL_WANT_LIBLTDL ? "yes" : "no";
|
|
|
|
mpirun_prefix_by_default = ORTE_WANT_ORTERUN_PREFIX_BY_DEFAULT ? "yes" : "no";
|
|
|
|
sparse_groups = OMPI_GROUP_SPARSE ? "yes" : "no";
|
|
|
|
have_mpi_io = OMPI_PROVIDE_MPI_FILE_INTERFACE ? "yes" : "no";
|
|
|
|
wtime_support = OPAL_TIMER_USEC_NATIVE ? "native" : "gettimeofday";
|
|
|
|
symbol_visibility = OPAL_C_HAVE_VISIBILITY ? "yes" : "no";
|
2011-09-13 04:39:14 +04:00
|
|
|
hwloc_support = OPAL_HAVE_HWLOC ? "yes" : "no";
|
2009-07-01 01:41:28 +04:00
|
|
|
|
|
|
|
/* setup strings that require allocation */
|
|
|
|
if (OMPI_WANT_F77_BINDINGS) {
|
|
|
|
asprintf(&f77, "yes (%s)",
|
|
|
|
(OPAL_HAVE_WEAK_SYMBOLS ? "all" :
|
|
|
|
(OMPI_F77_CAPS ? "caps" :
|
|
|
|
(OMPI_F77_PLAIN ? "lower case" :
|
|
|
|
(OMPI_F77_SINGLE_UNDERSCORE ? "single underscore" : "double underscore")))));
|
|
|
|
} else {
|
|
|
|
f77 = strdup("no");
|
|
|
|
}
|
|
|
|
|
2011-03-19 00:36:35 +03:00
|
|
|
if (OPAL_HAVE_SOLARIS_THREADS || OPAL_HAVE_POSIX_THREADS) { /* should just test OPAL_HAVE_THREADS */
|
|
|
|
asprintf(&threads, "%s (MPI_THREAD_MULTIPLE: %s, progress: %s)", OPAL_HAVE_SOLARIS_THREADS ? "solaris" :
|
|
|
|
(OPAL_HAVE_POSIX_THREADS ? "posix" : "type unknown"), /* "type unknown" can presumably never happen */
|
2010-03-17 02:10:50 +03:00
|
|
|
OMPI_ENABLE_THREAD_MULTIPLE ? "yes" : "no",
|
Update libevent to the 2.0 series, currently at 2.0.7rc. We will update to their final release when it becomes available. Currently known errors exist in unused portions of the libevent code. This revision passes the IBM test suite on a Linux machine and on a standalone Mac.
This is a fairly intrusive change, but outside of the moving of opal/event to opal/mca/event, the only changes involved (a) changing all calls to opal_event functions to reflect the new framework instead, and (b) ensuring that all opal_event_t objects are properly constructed since they are now true opal_objects.
Note: Shiqing has just returned from vacation and has not yet had a chance to complete the Windows integration. Thus, this commit almost certainly breaks Windows support on the trunk. However, I want this to have a chance to soak for as long as possible before I become less available a week from today (going to be at a class for 5 days, and thus will only be sparingly available) so we can find and fix any problems.
Biggest change is moving the libevent code from opal/event to a new opal/mca/event framework. This was done to make it much easier to update libevent in the future. New versions can be inserted as a new component and tested in parallel with the current version until validated, then we can remove the earlier version if we so choose. This is a statically built framework ala installdirs, so only one component will build at a time. There is no selection logic - the sole compiled component simply loads its function pointers into the opal_event struct.
I have gone thru the code base and converted all the libevent calls I could find. However, I cannot compile nor test every environment. It is therefore quite likely that errors remain in the system. Please keep an eye open for two things:
1. compile-time errors: these will be obvious as calls to the old functions (e.g., opal_evtimer_new) must be replaced by the new framework APIs (e.g., opal_event.evtimer_new)
2. run-time errors: these will likely show up as segfaults due to missing constructors on opal_event_t objects. It appears that it became a typical practice for people to "init" an opal_event_t by simply using memset to zero it out. This will no longer work - you must either OBJ_NEW or OBJ_CONSTRUCT an opal_event_t. I tried to catch these cases, but may have missed some. Believe me, you'll know when you hit it.
There is also the issue of the new libevent "no recursion" behavior. As I described on a recent email, we will have to discuss this and figure out what, if anything, we need to do.
This commit was SVN r23925.
2010-10-24 22:35:54 +04:00
|
|
|
OMPI_ENABLE_PROGRESS_THREADS ? "yes" : "no");
|
2009-07-01 01:41:28 +04:00
|
|
|
} else {
|
|
|
|
threads = strdup("no");
|
|
|
|
}
|
|
|
|
|
|
|
|
asprintf(&ft_support, "%s (checkpoint thread: %s)",
|
|
|
|
OPAL_ENABLE_FT ? "yes" : "no", OPAL_ENABLE_FT_THREAD ? "yes" : "no");;
|
A number of C/R enhancements per RFC below:
http://www.open-mpi.org/community/lists/devel/2010/07/8240.php
Documentation:
http://osl.iu.edu/research/ft/
Major Changes:
--------------
* Added C/R-enabled Debugging support.
Enabled with the --enable-crdebug flag. See the following website for more information:
http://osl.iu.edu/research/ft/crdebug/
* Added Stable Storage (SStore) framework for checkpoint storage
* 'central' component does a direct to central storage save
* 'stage' component stages checkpoints to central storage while the application continues execution.
* 'stage' supports offline compression of checkpoints before moving (sstore_stage_compress)
* 'stage' supports local caching of checkpoints to improve automatic recovery (sstore_stage_caching)
* Added Compression (compress) framework to support
* Add two new ErrMgr recovery policies
* {{{crmig}}} C/R Process Migration
* {{{autor}}} C/R Automatic Recovery
* Added the {{{ompi-migrate}}} command line tool to support the {{{crmig}}} ErrMgr component
* Added CR MPI Ext functions (enable them with {{{--enable-mpi-ext=cr}}} configure option)
* {{{OMPI_CR_Checkpoint}}} (Fixes trac:2342)
* {{{OMPI_CR_Restart}}}
* {{{OMPI_CR_Migrate}}} (may need some more work for mapping rules)
* {{{OMPI_CR_INC_register_callback}}} (Fixes trac:2192)
* {{{OMPI_CR_Quiesce_start}}}
* {{{OMPI_CR_Quiesce_checkpoint}}}
* {{{OMPI_CR_Quiesce_end}}}
* {{{OMPI_CR_self_register_checkpoint_callback}}}
* {{{OMPI_CR_self_register_restart_callback}}}
* {{{OMPI_CR_self_register_continue_callback}}}
* The ErrMgr predicted_fault() interface has been changed to take an opal_list_t of ErrMgr defined types. This will allow us to better support a wider range of fault prediction services in the future.
* Add a progress meter to:
* FileM rsh (filem_rsh_process_meter)
* SnapC full (snapc_full_progress_meter)
* SStore stage (sstore_stage_progress_meter)
* Added 2 new command line options to ompi-restart
* --showme : Display the full command line that would have been exec'ed.
* --mpirun_opts : Command line options to pass directly to mpirun. (Fixes trac:2413)
* Deprecated some MCA params:
* crs_base_snapshot_dir deprecated, use sstore_stage_local_snapshot_dir
* snapc_base_global_snapshot_dir deprecated, use sstore_base_global_snapshot_dir
* snapc_base_global_shared deprecated, use sstore_stage_global_is_shared
* snapc_base_store_in_place deprecated, replaced with different components of SStore
* snapc_base_global_snapshot_ref deprecated, use sstore_base_global_snapshot_ref
* snapc_base_establish_global_snapshot_dir deprecated, never well supported
* snapc_full_skip_filem deprecated, use sstore_stage_skip_filem
Minor Changes:
--------------
* Fixes trac:1924 : {{{ompi-restart}}} now recognizes path prefixed checkpoint handles and does the right thing.
* Fixes trac:2097 : {{{ompi-info}}} should now report all available CRS components
* Fixes trac:2161 : Manual checkpoint movement. A user can 'mv' a checkpoint directory from the original location to another and still restart from it.
* Fixes trac:2208 : Honor various TMPDIR varaibles instead of forcing {{{/tmp}}}
* Move {{{ompi_cr_continue_like_restart}}} to {{{orte_cr_continue_like_restart}}} to be more flexible in where this should be set.
* opal_crs_base_metadata_write* functions have been moved to SStore to support a wider range of metadata handling functionality.
* Cleanup the CRS framework and components to work with the SStore framework.
* Cleanup the SnapC framework and components to work with the SStore framework (cleans up these code paths considerably).
* Add 'quiesce' hook to CRCP for a future enhancement.
* We now require a BLCR version that supports {{{cr_request_file()}}} or {{{cr_request_checkpoint()}}} in order to make the code more maintainable. Note that {{{cr_request_file}}} has been deprecated since 0.7.0, so we prefer to use {{{cr_request_checkpoint()}}}.
* Add optional application level INC callbacks (registered through the CR MPI Ext interface).
* Increase the {{{opal_cr_thread_sleep_wait}}} parameter to 1000 microseconds to make the C/R thread less aggressive.
* {{{opal-restart}}} now looks for cache directories before falling back on stable storage when asked.
* {{{opal-restart}}} also support local decompression before restarting
* {{{orte-checkpoint}}} now uses the SStore framework to work with the metadata
* {{{orte-restart}}} now uses the SStore framework to work with the metadata
* Remove the {{{orte-restart}}} preload option. This was removed since the user only needs to select the 'stage' component in order to support this functionality.
* Since the '-am' parameter is saved in the metadata, {{{ompi-restart}}} no longer hard codes {{{-am ft-enable-cr}}}.
* Fix {{{hnp}}} ErrMgr so that if a previous component in the stack has 'fixed' the problem, then it should be skipped.
* Make sure to decrement the number of 'num_local_procs' in the orted when one goes away.
* odls now checks the SStore framework to see if it needs to load any checkpoint files before launching (to support 'stage'). This separates the SStore logic from the --preload-[binary|files] options.
* Add unique IDs to the named pipes established between the orted and the app in SnapC. This is to better support migration and automatic recovery activities.
* Improve the checks for 'already checkpointing' error path.
* A a recovery output timer, to show how long it takes to restart a job
* Do a better job of cleaning up the old session directory on restart.
* Add a local module to the autor and crmig ErrMgr components. These small modules prevent the 'orted' component from attempting a local recovery (Which does not work for MPI apps at the moment)
* Add a fix for bounding the checkpointable region between MPI_Init and MPI_Finalize.
This commit was SVN r23587.
The following Trac tickets were found above:
Ticket 1924 --> https://svn.open-mpi.org/trac/ompi/ticket/1924
Ticket 2097 --> https://svn.open-mpi.org/trac/ompi/ticket/2097
Ticket 2161 --> https://svn.open-mpi.org/trac/ompi/ticket/2161
Ticket 2192 --> https://svn.open-mpi.org/trac/ompi/ticket/2192
Ticket 2208 --> https://svn.open-mpi.org/trac/ompi/ticket/2208
Ticket 2342 --> https://svn.open-mpi.org/trac/ompi/ticket/2342
Ticket 2413 --> https://svn.open-mpi.org/trac/ompi/ticket/2413
2010-08-11 00:51:11 +04:00
|
|
|
|
|
|
|
asprintf(&crdebug_support, "%s",
|
|
|
|
OPAL_ENABLE_CRDEBUG ? "yes" : "no");
|
|
|
|
|
2009-07-01 01:41:28 +04:00
|
|
|
/* output values */
|
2010-11-13 02:22:11 +03:00
|
|
|
ompi_info_out("Configured by", "config:user", OPAL_CONFIGURE_USER);
|
|
|
|
ompi_info_out("Configured on", "config:timestamp", OPAL_CONFIGURE_DATE);
|
|
|
|
ompi_info_out("Configure host", "config:host", OPAL_CONFIGURE_HOST);
|
2009-07-01 01:41:28 +04:00
|
|
|
|
|
|
|
ompi_info_out("Built by", "build:user", OMPI_BUILD_USER);
|
|
|
|
ompi_info_out("Built on", "build:timestamp", OMPI_BUILD_DATE);
|
|
|
|
ompi_info_out("Built host", "build:host", OMPI_BUILD_HOST);
|
|
|
|
|
|
|
|
ompi_info_out("C bindings", "bindings:c", "yes");
|
|
|
|
ompi_info_out("C++ bindings", "bindings:cxx", cxx);
|
|
|
|
ompi_info_out("Fortran77 bindings", "bindings:f77", f77);
|
|
|
|
free(f77);
|
|
|
|
ompi_info_out("Fortran90 bindings", "bindings:f90", f90);
|
|
|
|
ompi_info_out("Fortran90 bindings size", "bindings:f90:size",
|
|
|
|
OMPI_WANT_F90_BINDINGS ? f90_size : "na");
|
|
|
|
|
|
|
|
ompi_info_out("C compiler", "compiler:c:command", OPAL_CC);
|
|
|
|
ompi_info_out("C compiler absolute", "compiler:c:absolute", OPAL_CC_ABSOLUTE);
|
|
|
|
ompi_info_out("C compiler family name", "compiler:c:familyname", _STRINGIFY(OPAL_BUILD_PLATFORM_COMPILER_FAMILYNAME));
|
|
|
|
ompi_info_out("C compiler version", "compiler:c:version", _STRINGIFY(OPAL_BUILD_PLATFORM_COMPILER_VERSION_STR));
|
|
|
|
|
|
|
|
if (want_all) {
|
|
|
|
ompi_info_out_int("C char size", "compiler:c:sizeof:char", sizeof(char));
|
2009-08-04 15:54:01 +04:00
|
|
|
/* JMS: should be fixed in MPI-2.2 to differentiate between C
|
|
|
|
_Bool and C++ bool. For the moment, the code base assumes
|
|
|
|
that they are the same. Because of opal_config_bottom.h,
|
|
|
|
we can sizeof(bool) here, so we might as well -- even
|
|
|
|
though this technically isn't right. This should be fixed
|
|
|
|
when we update to MPI-2.2. See below for note about C++
|
|
|
|
bool alignment. */
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("C bool size", "compiler:c:sizeof:bool", sizeof(bool));
|
|
|
|
ompi_info_out_int("C short size", "compiler:c:sizeof:short", sizeof(short));
|
|
|
|
ompi_info_out_int("C int size", "compiler:c:sizeof:int", sizeof(int));
|
|
|
|
ompi_info_out_int("C long size", "compiler:c:sizeof:long", sizeof(long));
|
|
|
|
ompi_info_out_int("C float size", "compiler:c:sizeof:float", sizeof(float));
|
|
|
|
ompi_info_out_int("C double size", "compiler:c:sizeof:double", sizeof(double));
|
|
|
|
ompi_info_out_int("C pointer size", "compiler:c:sizeof:pointer", sizeof(void *));
|
|
|
|
ompi_info_out_int("C char align", "compiler:c:align:char", OPAL_ALIGNMENT_CHAR);
|
2009-08-04 15:54:01 +04:00
|
|
|
#if OMPI_WANT_CXX_BINDINGS
|
|
|
|
/* JMS: See above for note about C++ bool size. We don't have
|
|
|
|
the bool alignment the way configure currently runs -- need
|
|
|
|
to clean this up when we update for MPI-2.2. */
|
- Split the datatype engine into two parts: an MPI specific part in
OMPI
and a language agnostic part in OPAL. The convertor is completely
moved into OPAL. This offers several benefits as described in RFC
http://www.open-mpi.org/community/lists/devel/2009/07/6387.php
namely:
- Fewer basic types (int* and float* types, boolean and wchar
- Fixing naming scheme to ompi-nomenclature.
- Usability outside of the ompi-layer.
- Due to the fixed nature of simple opal types, their information is
completely
known at compile time and therefore constified
- With fewer datatypes (22), the actual sizes of bit-field types may be
reduced
from 64 to 32 bits, allowing reorganizing the opal_datatype
structure, eliminating holes and keeping data required in convertor
(upon send/recv) in one cacheline...
This has implications to the convertor-datastructure and other parts
of the code.
- Several performance tests have been run, the netpipe latency does not
change with
this patch on Linux/x86-64 on the smoky cluster.
- Extensive tests have been done to verify correctness (no new
regressions) using:
1. mpi_test_suite on linux/x86-64 using clean ompi-trunk and
ompi-ddt:
a. running both trunk and ompi-ddt resulted in no differences
(except for MPI_SHORT_INT and MPI_TYPE_MIX_LB_UB do now run
correctly).
b. with --enable-memchecker and running under valgrind (one buglet
when run with static found in test-suite, commited)
2. ibm testsuite on linux/x86-64 using clean ompi-trunk and ompi-ddt:
all passed (except for the dynamic/ tests failed!! as trunk/MTT)
3. compilation and usage of HDF5 tests on Jaguar using PGI and
PathScale compilers.
4. compilation and usage on Scicortex.
- Please note, that for the heterogeneous case, (-m32 compiled
binaries/ompi), neither
ompi-trunk, nor ompi-ddt branch would successfully launch.
This commit was SVN r21641.
2009-07-13 08:56:31 +04:00
|
|
|
ompi_info_out_int("C bool align", "compiler:c:align:bool", OPAL_ALIGNMENT_CXX_BOOL);
|
2009-08-04 15:54:01 +04:00
|
|
|
#else
|
|
|
|
ompi_info_out("C bool align", "compiler:c:align:bool", "skipped");
|
|
|
|
#endif
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("C int align", "compiler:c:align:int", OPAL_ALIGNMENT_INT);
|
- Split the datatype engine into two parts: an MPI specific part in
OMPI
and a language agnostic part in OPAL. The convertor is completely
moved into OPAL. This offers several benefits as described in RFC
http://www.open-mpi.org/community/lists/devel/2009/07/6387.php
namely:
- Fewer basic types (int* and float* types, boolean and wchar
- Fixing naming scheme to ompi-nomenclature.
- Usability outside of the ompi-layer.
- Due to the fixed nature of simple opal types, their information is
completely
known at compile time and therefore constified
- With fewer datatypes (22), the actual sizes of bit-field types may be
reduced
from 64 to 32 bits, allowing reorganizing the opal_datatype
structure, eliminating holes and keeping data required in convertor
(upon send/recv) in one cacheline...
This has implications to the convertor-datastructure and other parts
of the code.
- Several performance tests have been run, the netpipe latency does not
change with
this patch on Linux/x86-64 on the smoky cluster.
- Extensive tests have been done to verify correctness (no new
regressions) using:
1. mpi_test_suite on linux/x86-64 using clean ompi-trunk and
ompi-ddt:
a. running both trunk and ompi-ddt resulted in no differences
(except for MPI_SHORT_INT and MPI_TYPE_MIX_LB_UB do now run
correctly).
b. with --enable-memchecker and running under valgrind (one buglet
when run with static found in test-suite, commited)
2. ibm testsuite on linux/x86-64 using clean ompi-trunk and ompi-ddt:
all passed (except for the dynamic/ tests failed!! as trunk/MTT)
3. compilation and usage of HDF5 tests on Jaguar using PGI and
PathScale compilers.
4. compilation and usage on Scicortex.
- Please note, that for the heterogeneous case, (-m32 compiled
binaries/ompi), neither
ompi-trunk, nor ompi-ddt branch would successfully launch.
This commit was SVN r21641.
2009-07-13 08:56:31 +04:00
|
|
|
ompi_info_out_int("C float align", "compiler:c:align:float", OPAL_ALIGNMENT_FLOAT);
|
|
|
|
ompi_info_out_int("C double align", "compiler:c:align:double", OPAL_ALIGNMENT_DOUBLE);
|
2009-07-01 01:41:28 +04:00
|
|
|
}
|
2009-08-04 15:54:01 +04:00
|
|
|
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out("C++ compiler", "compiler:cxx:command", OMPI_CXX);
|
|
|
|
ompi_info_out("C++ compiler absolute", "compiler:cxx:absolute", OMPI_CXX_ABSOLUTE);
|
|
|
|
ompi_info_out("Fortran77 compiler", "compiler:f77:command", OMPI_F77);
|
|
|
|
ompi_info_out("Fortran77 compiler abs", "compiler:f77:absolute",
|
|
|
|
OMPI_F77_ABSOLUTE);
|
|
|
|
ompi_info_out("Fortran90 compiler", "compiler:f90:command", OMPI_F90);
|
|
|
|
ompi_info_out("Fortran90 compiler abs", "compiler:f90:absolute",
|
|
|
|
OMPI_F90_ABSOLUTE);
|
|
|
|
|
|
|
|
if (want_all) {
|
|
|
|
|
|
|
|
/* Will always have the size of Fortran integer */
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort integer size", "compiler:fortran:sizeof:integer",
|
|
|
|
OMPI_SIZEOF_FORTRAN_INTEGER);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort logical size", "compiler:fortran:sizeof:logical",
|
|
|
|
OMPI_SIZEOF_FORTRAN_LOGICAL);
|
|
|
|
ompi_info_out_int("Fort logical value true", "compiler:fortran:value:true",
|
|
|
|
OMPI_FORTRAN_VALUE_TRUE);
|
|
|
|
|
|
|
|
|
|
|
|
/* May or may not have the other Fortran sizes */
|
|
|
|
|
|
|
|
if (OMPI_WANT_F77_BINDINGS || OMPI_WANT_F90_BINDINGS) {
|
|
|
|
ompi_info_out("Fort have integer1", "compiler:fortran:have:integer1",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER1 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have integer2", "compiler:fortran:have:integer2",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER2 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have integer4", "compiler:fortran:have:integer4",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER4 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have integer8", "compiler:fortran:have:integer8",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER8 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have integer16", "compiler:fortran:have:integer16",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER16 ? "yes" : "no");
|
|
|
|
|
|
|
|
ompi_info_out("Fort have real4", "compiler:fortran:have:real4",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL4 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have real8", "compiler:fortran:have:real8",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL8 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have real16", "compiler:fortran:have:real16",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL16 && OMPI_REAL16_MATCHES_C ? "yes" : "no");
|
|
|
|
|
|
|
|
ompi_info_out("Fort have complex8", "compiler:fortran:have:complex8",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX8 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have complex16", "compiler:fortran:have:complex16",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX16 ? "yes" : "no");
|
|
|
|
ompi_info_out("Fort have complex32", "compiler:fortran:have:complex32",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX32 && OMPI_REAL16_MATCHES_C ? "yes" : "no");
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort integer1 size", "compiler:fortran:sizeof:integer1",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER1 ? OMPI_SIZEOF_FORTRAN_INTEGER1 : -1);
|
|
|
|
ompi_info_out_int("Fort integer2 size", "compiler:fortran:sizeof:integer2",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER2 ? OMPI_SIZEOF_FORTRAN_INTEGER2 : -1);
|
|
|
|
ompi_info_out_int("Fort integer4 size", "compiler:fortran:sizeof:integer4",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER4 ? OMPI_SIZEOF_FORTRAN_INTEGER4 : -1);
|
|
|
|
ompi_info_out_int("Fort integer8 size", "compiler:fortran:sizeof:integer8",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER8 ? OMPI_SIZEOF_FORTRAN_INTEGER8 : -1);
|
2010-08-09 20:56:32 +04:00
|
|
|
ompi_info_out_int("Fort integer16 size", "compiler:fortran:sizeof:integer16",
|
2009-07-01 01:41:28 +04:00
|
|
|
OMPI_HAVE_FORTRAN_INTEGER16 ? OMPI_SIZEOF_FORTRAN_INTEGER16 : -1);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort real size", "compiler:fortran:sizeof:real",
|
|
|
|
OMPI_SIZEOF_FORTRAN_REAL);
|
|
|
|
ompi_info_out_int("Fort real4 size", "compiler:fortran:sizeof:real4",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL4 ? OMPI_SIZEOF_FORTRAN_REAL4 : -1);
|
|
|
|
ompi_info_out_int("Fort real8 size", "compiler:fortran:sizeof:real8",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL8 ? OMPI_SIZEOF_FORTRAN_REAL8 : -1);
|
|
|
|
ompi_info_out_int("Fort real16 size", "compiler:fortran:sizeof:real17",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL16 ? OMPI_SIZEOF_FORTRAN_REAL16 : -1);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort dbl prec size",
|
|
|
|
"compiler:fortran:sizeof:double_precision",
|
2010-08-09 20:56:32 +04:00
|
|
|
OMPI_SIZEOF_FORTRAN_DOUBLE_PRECISION);
|
2009-07-01 01:41:28 +04:00
|
|
|
|
|
|
|
ompi_info_out_int("Fort cplx size", "compiler:fortran:sizeof:complex",
|
2010-08-09 20:56:32 +04:00
|
|
|
OMPI_SIZEOF_FORTRAN_COMPLEX);
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("Fort dbl cplx size",
|
|
|
|
"compiler:fortran:sizeof:double_complex",
|
2010-08-11 00:33:50 +04:00
|
|
|
OMPI_HAVE_FORTRAN_DOUBLE_COMPLEX ? OMPI_SIZEOF_FORTRAN_DOUBLE_COMPLEX : -1);
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("Fort cplx8 size", "compiler:fortran:sizeof:complex8",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX8 ? OMPI_SIZEOF_FORTRAN_COMPLEX8 : -1);
|
|
|
|
ompi_info_out_int("Fort cplx16 size", "compiler:fortran:sizeof:complex16",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX16 ? OMPI_SIZEOF_FORTRAN_COMPLEX16 : -1);
|
|
|
|
ompi_info_out_int("Fort cplx32 size", "compiler:fortran:sizeof:complex32",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX32 ? OMPI_SIZEOF_FORTRAN_COMPLEX32 : -1);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort integer align", "compiler:fortran:align:integer",
|
|
|
|
OMPI_ALIGNMENT_FORTRAN_INTEGER);
|
|
|
|
ompi_info_out_int("Fort integer1 align", "compiler:fortran:align:integer1",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER1 ? OMPI_ALIGNMENT_FORTRAN_INTEGER1 : -1);
|
|
|
|
ompi_info_out_int("Fort integer2 align", "compiler:fortran:align:integer2",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER2 ? OMPI_ALIGNMENT_FORTRAN_INTEGER2 : -1);
|
|
|
|
ompi_info_out_int("Fort integer4 align", "compiler:fortran:align:integer4",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER4 ? OMPI_ALIGNMENT_FORTRAN_INTEGER4 : -1);
|
|
|
|
ompi_info_out_int("Fort integer8 align", "compiler:fortran:align:integer8",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER8 ? OMPI_ALIGNMENT_FORTRAN_INTEGER8 : -1);
|
|
|
|
ompi_info_out_int("Fort integer16 align", "compiler:fortran:align:integer16",
|
|
|
|
OMPI_HAVE_FORTRAN_INTEGER16 ? OMPI_ALIGNMENT_FORTRAN_INTEGER16 : -1);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort real align", "compiler:fortran:align:real",
|
|
|
|
OMPI_ALIGNMENT_FORTRAN_REAL);
|
|
|
|
ompi_info_out_int("Fort real4 align", "compiler:fortran:align:real4",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL4 ? OMPI_ALIGNMENT_FORTRAN_REAL4 : -1);
|
|
|
|
ompi_info_out_int("Fort real8 align", "compiler:fortran:align:real8",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL8 ? OMPI_ALIGNMENT_FORTRAN_REAL8 : -1);
|
|
|
|
ompi_info_out_int("Fort real16 align", "compiler:fortran:align:real16",
|
|
|
|
OMPI_HAVE_FORTRAN_REAL16 ? OMPI_ALIGNMENT_FORTRAN_REAL16 : -1);
|
|
|
|
|
|
|
|
ompi_info_out_int("Fort dbl prec align",
|
|
|
|
"compiler:fortran:align:double_precision",
|
2010-08-09 20:56:32 +04:00
|
|
|
OMPI_ALIGNMENT_FORTRAN_DOUBLE_PRECISION);
|
2009-07-01 01:41:28 +04:00
|
|
|
|
|
|
|
ompi_info_out_int("Fort cplx align", "compiler:fortran:align:complex",
|
2010-08-09 20:56:32 +04:00
|
|
|
OMPI_ALIGNMENT_FORTRAN_COMPLEX);
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("Fort dbl cplx align",
|
|
|
|
"compiler:fortran:align:double_complex",
|
2010-08-11 00:33:50 +04:00
|
|
|
OMPI_HAVE_FORTRAN_DOUBLE_COMPLEX ? OMPI_ALIGNMENT_FORTRAN_DOUBLE_COMPLEX : -1);
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("Fort cplx8 align", "compiler:fortran:align:complex8",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX8 ? OMPI_ALIGNMENT_FORTRAN_COMPLEX8 : -1);
|
|
|
|
ompi_info_out_int("Fort cplx16 align", "compiler:fortran:align:complex16",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX16 ? OMPI_ALIGNMENT_FORTRAN_COMPLEX16 : -1);
|
|
|
|
ompi_info_out_int("Fort cplx32 align", "compiler:fortran:align:complex32",
|
|
|
|
OMPI_HAVE_FORTRAN_COMPLEX32 ? OMPI_ALIGNMENT_FORTRAN_COMPLEX32 : -1);
|
|
|
|
|
|
|
|
} else {
|
|
|
|
ompi_info_out("Fort real size", "compiler:fortran:sizeof:real", "skipped");
|
|
|
|
ompi_info_out("Fort dbl prec size",
|
|
|
|
"compiler:fortran:sizeof:double_precision", "skipped");
|
|
|
|
ompi_info_out("Fort cplx size", "compiler:fortran:sizeof:complex", "skipped");
|
|
|
|
ompi_info_out("Fort dbl cplx size",
|
|
|
|
"compiler:fortran:sizeof:double_complex", "skipped");
|
|
|
|
|
|
|
|
ompi_info_out("Fort integer align", "compiler:fortran:align:integer", "skipped");
|
|
|
|
ompi_info_out("Fort real align", "compiler:fortran:align:real", "skipped");
|
|
|
|
ompi_info_out("Fort dbl prec align",
|
|
|
|
"compiler:fortran:align:double_precision","skipped");
|
|
|
|
ompi_info_out("Fort cplx align", "compiler:fortran:align:complex", "skipped");
|
|
|
|
ompi_info_out("Fort dbl cplx align",
|
|
|
|
"compiler:fortran:align:double_complex", "skipped");
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
ompi_info_out("C profiling", "option:profiling:c", cprofiling);
|
|
|
|
ompi_info_out("C++ profiling", "option:profiling:cxx", cxxprofiling);
|
|
|
|
ompi_info_out("Fortran77 profiling", "option:profiling:f77", f77profiling);
|
|
|
|
ompi_info_out("Fortran90 profiling", "option:profiling:f90", f90profiling);
|
|
|
|
|
|
|
|
ompi_info_out("C++ exceptions", "option:cxx_exceptions", cxxexceptions);
|
|
|
|
ompi_info_out("Thread support", "option:threads", threads);
|
|
|
|
free(threads);
|
|
|
|
ompi_info_out("Sparse Groups", "option:sparse:groups", sparse_groups);
|
|
|
|
|
|
|
|
if (want_all) {
|
|
|
|
|
|
|
|
/* Don't display the build CPPFLAGS or CXXCPPFLAGS because they're
|
|
|
|
* just -I$(top_srcdir)/include, etc. Hence, they're a) boring,
|
|
|
|
* and c) specific for ompi_info.
|
|
|
|
*/
|
|
|
|
|
|
|
|
ompi_info_out("Build CFLAGS", "option:build:cflags", OMPI_BUILD_CFLAGS);
|
|
|
|
ompi_info_out("Build CXXFLAGS", "option:build:cxxflags", OMPI_BUILD_CXXFLAGS);
|
|
|
|
ompi_info_out("Build FFLAGS", "option:build:fflags", OMPI_BUILD_FFLAGS);
|
|
|
|
ompi_info_out("Build FCFLAGS", "option:build:fcflags", OMPI_BUILD_FCFLAGS);
|
|
|
|
ompi_info_out("Build LDFLAGS", "option:build:ldflags", OMPI_BUILD_LDFLAGS);
|
|
|
|
ompi_info_out("Build LIBS", "option:build:libs", OMPI_BUILD_LIBS);
|
|
|
|
|
|
|
|
ompi_info_out("Wrapper extra CFLAGS", "option:wrapper:extra_cflags",
|
|
|
|
WRAPPER_EXTRA_CFLAGS);
|
|
|
|
ompi_info_out("Wrapper extra CXXFLAGS", "option:wrapper:extra_cxxflags",
|
|
|
|
WRAPPER_EXTRA_CXXFLAGS);
|
|
|
|
ompi_info_out("Wrapper extra FFLAGS", "option:wrapper:extra_fflags",
|
|
|
|
WRAPPER_EXTRA_FFLAGS);
|
|
|
|
ompi_info_out("Wrapper extra FCFLAGS", "option:wrapper:extra_fcflags",
|
|
|
|
WRAPPER_EXTRA_FCFLAGS);
|
|
|
|
ompi_info_out("Wrapper extra LDFLAGS", "option:wrapper:extra_ldflags",
|
|
|
|
WRAPPER_EXTRA_LDFLAGS);
|
|
|
|
ompi_info_out("Wrapper extra LIBS", "option:wrapper:extra_libs",
|
|
|
|
WRAPPER_EXTRA_LIBS);
|
|
|
|
}
|
|
|
|
|
|
|
|
ompi_info_out("Internal debug support", "option:debug", debug);
|
|
|
|
ompi_info_out("MPI interface warnings", "option:mpi-interface-warning", mpi_interface_warning);
|
|
|
|
ompi_info_out("MPI parameter check", "option:mpi-param-check", paramcheck);
|
|
|
|
ompi_info_out("Memory profiling support", "option:mem-profile", memprofile);
|
|
|
|
ompi_info_out("Memory debugging support", "option:mem-debug", memdebug);
|
2010-05-18 03:08:56 +04:00
|
|
|
ompi_info_out("Unusual events notif.", "option:log-event", log_event);
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out("libltdl support", "option:dlopen", want_libltdl);
|
|
|
|
ompi_info_out("Heterogeneous support", "options:heterogeneous", heterogeneous);
|
|
|
|
ompi_info_out("mpirun default --prefix", "mpirun:prefix_by_default",
|
|
|
|
mpirun_prefix_by_default);
|
|
|
|
ompi_info_out("MPI I/O support", "options:mpi-io", have_mpi_io);
|
|
|
|
ompi_info_out("MPI_WTIME support", "options:mpi-wtime", wtime_support);
|
2009-07-29 18:22:46 +04:00
|
|
|
ompi_info_out("Symbol vis. support", "options:visibility", symbol_visibility);
|
2011-09-13 04:39:14 +04:00
|
|
|
ompi_info_out("Hwloc support", "options:hwloc", hwloc_support);
|
2009-07-01 01:41:28 +04:00
|
|
|
|
2010-08-17 08:44:22 +04:00
|
|
|
ompi_info_out("MPI extensions", "options:mpi_ext", OMPI_MPIEXT_COMPONENTS);
|
2009-07-01 01:41:28 +04:00
|
|
|
|
|
|
|
ompi_info_out("FT Checkpoint support", "options:ft_support", ft_support);
|
|
|
|
free(ft_support);
|
A number of C/R enhancements per RFC below:
http://www.open-mpi.org/community/lists/devel/2010/07/8240.php
Documentation:
http://osl.iu.edu/research/ft/
Major Changes:
--------------
* Added C/R-enabled Debugging support.
Enabled with the --enable-crdebug flag. See the following website for more information:
http://osl.iu.edu/research/ft/crdebug/
* Added Stable Storage (SStore) framework for checkpoint storage
* 'central' component does a direct to central storage save
* 'stage' component stages checkpoints to central storage while the application continues execution.
* 'stage' supports offline compression of checkpoints before moving (sstore_stage_compress)
* 'stage' supports local caching of checkpoints to improve automatic recovery (sstore_stage_caching)
* Added Compression (compress) framework to support
* Add two new ErrMgr recovery policies
* {{{crmig}}} C/R Process Migration
* {{{autor}}} C/R Automatic Recovery
* Added the {{{ompi-migrate}}} command line tool to support the {{{crmig}}} ErrMgr component
* Added CR MPI Ext functions (enable them with {{{--enable-mpi-ext=cr}}} configure option)
* {{{OMPI_CR_Checkpoint}}} (Fixes trac:2342)
* {{{OMPI_CR_Restart}}}
* {{{OMPI_CR_Migrate}}} (may need some more work for mapping rules)
* {{{OMPI_CR_INC_register_callback}}} (Fixes trac:2192)
* {{{OMPI_CR_Quiesce_start}}}
* {{{OMPI_CR_Quiesce_checkpoint}}}
* {{{OMPI_CR_Quiesce_end}}}
* {{{OMPI_CR_self_register_checkpoint_callback}}}
* {{{OMPI_CR_self_register_restart_callback}}}
* {{{OMPI_CR_self_register_continue_callback}}}
* The ErrMgr predicted_fault() interface has been changed to take an opal_list_t of ErrMgr defined types. This will allow us to better support a wider range of fault prediction services in the future.
* Add a progress meter to:
* FileM rsh (filem_rsh_process_meter)
* SnapC full (snapc_full_progress_meter)
* SStore stage (sstore_stage_progress_meter)
* Added 2 new command line options to ompi-restart
* --showme : Display the full command line that would have been exec'ed.
* --mpirun_opts : Command line options to pass directly to mpirun. (Fixes trac:2413)
* Deprecated some MCA params:
* crs_base_snapshot_dir deprecated, use sstore_stage_local_snapshot_dir
* snapc_base_global_snapshot_dir deprecated, use sstore_base_global_snapshot_dir
* snapc_base_global_shared deprecated, use sstore_stage_global_is_shared
* snapc_base_store_in_place deprecated, replaced with different components of SStore
* snapc_base_global_snapshot_ref deprecated, use sstore_base_global_snapshot_ref
* snapc_base_establish_global_snapshot_dir deprecated, never well supported
* snapc_full_skip_filem deprecated, use sstore_stage_skip_filem
Minor Changes:
--------------
* Fixes trac:1924 : {{{ompi-restart}}} now recognizes path prefixed checkpoint handles and does the right thing.
* Fixes trac:2097 : {{{ompi-info}}} should now report all available CRS components
* Fixes trac:2161 : Manual checkpoint movement. A user can 'mv' a checkpoint directory from the original location to another and still restart from it.
* Fixes trac:2208 : Honor various TMPDIR varaibles instead of forcing {{{/tmp}}}
* Move {{{ompi_cr_continue_like_restart}}} to {{{orte_cr_continue_like_restart}}} to be more flexible in where this should be set.
* opal_crs_base_metadata_write* functions have been moved to SStore to support a wider range of metadata handling functionality.
* Cleanup the CRS framework and components to work with the SStore framework.
* Cleanup the SnapC framework and components to work with the SStore framework (cleans up these code paths considerably).
* Add 'quiesce' hook to CRCP for a future enhancement.
* We now require a BLCR version that supports {{{cr_request_file()}}} or {{{cr_request_checkpoint()}}} in order to make the code more maintainable. Note that {{{cr_request_file}}} has been deprecated since 0.7.0, so we prefer to use {{{cr_request_checkpoint()}}}.
* Add optional application level INC callbacks (registered through the CR MPI Ext interface).
* Increase the {{{opal_cr_thread_sleep_wait}}} parameter to 1000 microseconds to make the C/R thread less aggressive.
* {{{opal-restart}}} now looks for cache directories before falling back on stable storage when asked.
* {{{opal-restart}}} also support local decompression before restarting
* {{{orte-checkpoint}}} now uses the SStore framework to work with the metadata
* {{{orte-restart}}} now uses the SStore framework to work with the metadata
* Remove the {{{orte-restart}}} preload option. This was removed since the user only needs to select the 'stage' component in order to support this functionality.
* Since the '-am' parameter is saved in the metadata, {{{ompi-restart}}} no longer hard codes {{{-am ft-enable-cr}}}.
* Fix {{{hnp}}} ErrMgr so that if a previous component in the stack has 'fixed' the problem, then it should be skipped.
* Make sure to decrement the number of 'num_local_procs' in the orted when one goes away.
* odls now checks the SStore framework to see if it needs to load any checkpoint files before launching (to support 'stage'). This separates the SStore logic from the --preload-[binary|files] options.
* Add unique IDs to the named pipes established between the orted and the app in SnapC. This is to better support migration and automatic recovery activities.
* Improve the checks for 'already checkpointing' error path.
* A a recovery output timer, to show how long it takes to restart a job
* Do a better job of cleaning up the old session directory on restart.
* Add a local module to the autor and crmig ErrMgr components. These small modules prevent the 'orted' component from attempting a local recovery (Which does not work for MPI apps at the moment)
* Add a fix for bounding the checkpointable region between MPI_Init and MPI_Finalize.
This commit was SVN r23587.
The following Trac tickets were found above:
Ticket 1924 --> https://svn.open-mpi.org/trac/ompi/ticket/1924
Ticket 2097 --> https://svn.open-mpi.org/trac/ompi/ticket/2097
Ticket 2161 --> https://svn.open-mpi.org/trac/ompi/ticket/2161
Ticket 2192 --> https://svn.open-mpi.org/trac/ompi/ticket/2192
Ticket 2208 --> https://svn.open-mpi.org/trac/ompi/ticket/2208
Ticket 2342 --> https://svn.open-mpi.org/trac/ompi/ticket/2342
Ticket 2413 --> https://svn.open-mpi.org/trac/ompi/ticket/2413
2010-08-11 00:51:11 +04:00
|
|
|
|
|
|
|
ompi_info_out("C/R Enabled Debugging", "options:crdebug_support", crdebug_support);
|
|
|
|
free(crdebug_support);
|
|
|
|
|
2009-07-01 01:41:28 +04:00
|
|
|
ompi_info_out_int("MPI_MAX_PROCESSOR_NAME", "options:mpi-max-processor-name",
|
|
|
|
MPI_MAX_PROCESSOR_NAME);
|
|
|
|
ompi_info_out_int("MPI_MAX_ERROR_STRING", "options:mpi-max-error-string",
|
|
|
|
MPI_MAX_ERROR_STRING);
|
|
|
|
ompi_info_out_int("MPI_MAX_OBJECT_NAME", "options:mpi-max-object-name",
|
|
|
|
MPI_MAX_OBJECT_NAME);
|
|
|
|
ompi_info_out_int("MPI_MAX_INFO_KEY", "options:mpi-max-info-key",
|
|
|
|
MPI_MAX_INFO_KEY);
|
|
|
|
ompi_info_out_int("MPI_MAX_INFO_VAL", "options:mpi-max-info-val",
|
|
|
|
MPI_MAX_INFO_VAL);
|
|
|
|
ompi_info_out_int("MPI_MAX_PORT_NAME", "options:mpi-max-port-name",
|
|
|
|
MPI_MAX_PORT_NAME);
|
|
|
|
#if OMPI_PROVIDE_MPI_FILE_INTERFACE
|
2009-07-07 19:28:34 +04:00
|
|
|
ompi_info_out_int("MPI_MAX_DATAREP_STRING", "options:mpi-max-datarep-string",
|
|
|
|
MPI_MAX_DATAREP_STRING);
|
2009-07-01 01:41:28 +04:00
|
|
|
#else
|
2009-07-07 19:28:34 +04:00
|
|
|
ompi_info_out("MPI_MAX_DATAREP_STRING", "options:mpi-max-datarep-string",
|
|
|
|
"IO interface not provided");
|
2009-07-01 01:41:28 +04:00
|
|
|
#endif
|
|
|
|
|
|
|
|
}
|