Add two new frameworks for sensing and predicting faults. This is just the bare-bones plumbing for now - will instantiate soon.

No ess modules reference these frameworks yet, so they are completely inactive.

This commit was SVN r21847.

orte/mca/fddp/Makefile.am

@@ -0,0 +1,34 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+# main library setup
+noinst_LTLIBRARIES = libmca_fddp.la
+libmca_fddp_la_SOURCES =
+
+# header setup
+nobase_orte_HEADERS =
+
+# local files
+headers = fddp.h
+
+libmca_fddp_la_SOURCES += $(headers)
+
+# Conditionally install the header files
+if WANT_INSTALL_HEADERS
+nobase_orte_HEADERS += $(headers)
+ortedir = $(includedir)/openmpi/orte/mca/fddp
+else
+ortedir = $(includedir)
+endif
+
+include base/Makefile.am
+
+distclean-local:
+	rm -f base/static-components.h

orte/mca/fddp/base/Makefile.am

@@ -0,0 +1,17 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+#
+# Additional copyrights may follow
+#
+# $HEADER$
+#
+
+headers += \
+        base/base.h
+
+libmca_fddp_la_SOURCES += \
+        base/fddp_base_close.c \
+        base/fddp_base_select.c \
+        base/fddp_base_open.c

orte/mca/fddp/base/base.h

@@ -0,0 +1,55 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+/** @file:
+ */
+
+#ifndef MCA_FDDP_BASE_H
+#define MCA_FDDP_BASE_H
+
+/*
+ * includes
+ */
+#include "orte_config.h"
+
+#include "opal/class/opal_list.h"
+#include "opal/mca/mca.h"
+
+#include "orte/mca/fddp/fddp.h"
+
+
+/*
+ * Global functions for MCA overall collective open and close
+ */
+BEGIN_C_DECLS
+
+/*
+ * function definitions
+ */
+ORTE_DECLSPEC    int orte_fddp_base_open(void);
+ORTE_DECLSPEC    int orte_fddp_base_select(void);
+ORTE_DECLSPEC    int orte_fddp_base_close(void);
+
+/*
+ * globals that might be needed
+ */
+
+ORTE_DECLSPEC extern int orte_fddp_base_output;
+ORTE_DECLSPEC extern bool mca_fddp_base_selected;
+ORTE_DECLSPEC extern opal_list_t mca_fddp_base_components_available;
+ORTE_DECLSPEC extern orte_fddp_base_component_t mca_fddp_base_selected_component;
+
+#if !ORTE_DISABLE_FULL_SUPPORT
+
+/* no base functions to protect at this time */
+
+#endif /* ORTE_DISABLE_FULL_SUPPORT */
+
+END_C_DECLS
+#endif

orte/mca/fddp/base/fddp_base_close.c

@@ -0,0 +1,38 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include <stdio.h>
+
+#include "opal/mca/mca.h"
+#include "opal/mca/base/base.h"
+#include "orte/mca/fddp/base/base.h"
+
+
+int orte_fddp_base_close(void)
+{
+    /* If we have a selected component and module, then finalize it */
+    
+    if (NULL != orte_fddp.finalize) {
+        orte_fddp.finalize();
+    }
+    
+    /* Close all remaining available components (may be one if this is a
+     OpenRTE program, or [possibly] multiple if this is ompi_info) */
+    
+    mca_base_components_close(orte_fddp_base_output, 
+                              &mca_fddp_base_components_available, NULL);
+    
+    /* All done */
+    
+    return ORTE_SUCCESS;
+}

orte/mca/fddp/base/fddp_base_open.c

@@ -0,0 +1,66 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/mca.h"
+#include "opal/util/output.h"
+#include "opal/mca/base/base.h"
+#include "opal/mca/base/mca_base_param.h"
+
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "orte/mca/fddp/base/base.h"
+
+
+/*
+ * The following file was created by configure.  It contains extern
+ * statements and the definition of an array of pointers to each
+ * component's public mca_base_component_t struct.
+ */
+
+#include "orte/mca/fddp/base/static-components.h"
+
+/*
+ * Global variables
+ */
+int orte_fddp_base_output = -1;
+orte_fddp_base_module_t orte_fddp;
+opal_list_t mca_fddp_base_components_available;
+orte_fddp_base_component_t mca_fddp_base_selected_component;
+
+
+/**
+ * Function for finding and opening either all MCA components, or the one
+ * that was specifically requested via a MCA parameter.
+ */
+int orte_fddp_base_open(void)
+{
+    /* Debugging / verbose output.  Always have stream open, with
+       verbose set by the mca open system... */
+    orte_fddp_base_output = opal_output_open(NULL);
+    
+    /* Open up all available components */
+
+    if (ORTE_SUCCESS !=
+        mca_base_components_open("fddp", orte_fddp_base_output,
+                                 mca_fddp_base_static_components,
+                                 &mca_fddp_base_components_available, true)) {
+        return ORTE_ERROR;
+    }
+
+    /* All done */
+
+    return ORTE_SUCCESS;
+}

orte/mca/fddp/base/fddp_base_select.c

@@ -0,0 +1,78 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/mca.h"
+#include "opal/mca/base/base.h"
+
+#include "orte/mca/fddp/base/base.h"
+
+
+/**
+ * Function for selecting one component from all those that are
+ * available.
+ */
+int orte_fddp_base_select(void)
+{
+    int ret, exit_status = ORTE_SUCCESS;
+    orte_fddp_base_component_t *best_component = NULL;
+    orte_fddp_base_module_t *best_module = NULL;
+    char *include_list = NULL;
+
+    /*
+     * Register the framework MCA param and look up include list
+     */
+    mca_base_param_reg_string_name("fddp", NULL,
+                                   "Which fddp component to use (empty = none)",
+                                   false, false,
+                                   NULL, &include_list);
+    
+    /* If we do not have any components to select this is ok. Just use the default
+     * "no-op" component and move on.
+     */
+    if( 0 >= opal_list_get_size(&mca_fddp_base_components_available) || NULL == include_list) { 
+        /* Close all components since none will be used */
+        mca_base_components_close(0, /* Pass 0 to keep this from closing the output handle */
+                                  &mca_fddp_base_components_available,
+                                  NULL);
+        goto cleanup;
+    }
+    
+    /*
+     * Select the best component
+     */
+    if( ORTE_SUCCESS != mca_base_select("fddp", orte_fddp_base_output,
+                                        &mca_fddp_base_components_available,
+                                        (mca_base_module_t **) &best_module,
+                                        (mca_base_component_t **) &best_component) ) {
+        /* It is okay if no component was selected - we just leave
+         * the orte_fddp module as the default
+         */
+        exit_status = ORTE_SUCCESS;
+        goto cleanup;
+    }
+
+    if (NULL != orte_fddp.init) {
+        /* if an init function is provided, use it */
+        if (ORTE_SUCCESS != (ret = orte_fddp.init()) ) {
+            exit_status = ret;
+            goto cleanup;
+        }
+    }
+
+    /* Save the winner */
+    orte_fddp = *best_module;
+
+ cleanup:
+    return exit_status;
+}

orte/mca/fddp/fddp.h

@@ -0,0 +1,81 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ *
+ * @file:
+ *
+ */
+
+#ifndef MCA_FDDP_H
+#define MCA_FDDP_H
+
+/*
+ * includes
+ */
+
+#include "orte_config.h"
+
+#include "opal/mca/mca.h"
+
+#include "orte/mca/sensor/sensor_types.h"
+
+BEGIN_C_DECLS
+
+/*
+ * Component functions - all MUST be provided!
+ */
+
+/* initialize the selected module */
+typedef int (*orte_fddp_base_module_init_fn_t)(void);
+    
+/* finalize the selected module */
+typedef int (*orte_fddp_base_module_finalize_fn_t)(void);
+
+typedef int (*orte_fddp_base_module_process_fn_t)(orte_sensor_data_t *data,
+                                                  int num_bins, uint8_t *failure_likelihood);
+    
+/*
+ * Ver 1.0
+ */
+struct orte_fddp_base_module_1_0_0_t {
+    orte_fddp_base_module_init_fn_t       init;
+    orte_fddp_base_module_finalize_fn_t   finalize;
+    orte_fddp_base_module_process_fn_t    process;
+};
+
+typedef struct orte_fddp_base_module_1_0_0_t orte_fddp_base_module_1_0_0_t;
+typedef orte_fddp_base_module_1_0_0_t orte_fddp_base_module_t;
+
+/*
+ * the standard component data structure
+ */
+struct orte_fddp_base_component_1_0_0_t {
+    mca_base_component_t base_version;
+    mca_base_component_data_t base_data;
+};
+typedef struct orte_fddp_base_component_1_0_0_t orte_fddp_base_component_1_0_0_t;
+typedef orte_fddp_base_component_1_0_0_t orte_fddp_base_component_t;
+
+
+
+/*
+ * Macro for use in components that are of type fddp v1.0.0
+ */
+#define ORTE_FDDP_BASE_VERSION_1_0_0 \
+  /* fddp v1.0 is chained to MCA v2.0 */ \
+  MCA_BASE_VERSION_2_0_0, \
+  /* fddp v1.0 */ \
+  "fddp", 1, 0, 0
+
+/* Global structure for accessing fddp functions
+ */
+ORTE_DECLSPEC extern orte_fddp_base_module_t orte_fddp;  /* holds selected module's function pointers */
+
+END_C_DECLS
+
+#endif /* MCA_FDDP_H */

orte/mca/fddp/trend/Makefile.am

@@ -0,0 +1,35 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+sources = \
+        fddp_trend.c \
+        fddp_trend.h \
+        fddp_trend_component.c
+
+# Make the output library in this directory, and name it either
+# mca_<type>_<name>.la (for DSO builds) or libmca_<type>_<name>.la
+# (for static builds).
+
+if OMPI_BUILD_fddp_trend_DSO
+component_noinst =
+component_install = mca_fddp_trend.la
+else
+component_noinst = libmca_fddp_trend.la
+component_install =
+endif
+
+mcacomponentdir = $(pkglibdir)
+mcacomponent_LTLIBRARIES = $(component_install)
+mca_fddp_trend_la_SOURCES = $(sources)
+mca_fddp_trend_la_LDFLAGS = -module -avoid-version
+
+noinst_LTLIBRARIES = $(component_noinst)
+libmca_fddp_trend_la_SOURCES =$(sources)
+libmca_fddp_trend_la_LDFLAGS = -module -avoid-version

orte/mca/fddp/trend/configure.params

@@ -0,0 +1,14 @@
+# -*- shell-script -*-
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved.
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+# Specific to this module
+
+PARAM_CONFIG_FILES="Makefile"

orte/mca/fddp/trend/fddp_trend.c

@@ -0,0 +1,69 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+#include "orte/types.h"
+
+#include <errno.h>
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif  /* HAVE_UNISTD_H */
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif  /* HAVE_STRING_H */
+#include <stdio.h>
+
+#include "opal/mca/base/mca_base_param.h"
+#include "opal/util/argv.h"
+#include "opal/class/opal_pointer_array.h"
+
+#include "orte/util/show_help.h"
+#include "orte/mca/errmgr/errmgr.h"
+#include "orte/mca/sensor/sensor_types.h"
+
+#include "fddp_trend.h"
+
+static int init(void);
+static int finalize(void);
+static int process(orte_sensor_data_t *data, int num_bins, uint8_t *failure_likelihood);
+
+orte_fddp_base_module_t orte_fddp_trend_module = {
+    init,
+    finalize,
+    process
+};
+
+static int init(void)
+{
+    return ORTE_SUCCESS;
+}
+
+static int finalize(void)
+{
+    return ORTE_SUCCESS;
+}
+
+static int process(orte_sensor_data_t *data, int num_bins, uint8_t *failure_likelihood)
+{
+    /* the failure likelihood in this model is just the trended value of the
+     * data itself, scaled appropriately
+     */
+    
+    /* using the sliding window, compute the trend of the data */
+    
+    /* for each point in future time, compute the predicted value of
+     * the sensor reading
+     */
+    
+    /* scale it by the provided scaling factors */
+    
+    return ORTE_SUCCESS;
+}

orte/mca/fddp/trend/fddp_trend.h

@@ -0,0 +1,35 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+/**
+ * @file
+ *
+ */
+#ifndef ORTE_FDDP_TREND_H
+#define ORTE_FDDP_TREND_H
+
+#include "orte_config.h"
+
+#include "orte/mca/fddp/fddp.h"
+
+BEGIN_C_DECLS
+
+struct orte_fddp_trend_component_t {
+    orte_fddp_base_component_t super;
+    int window_size;
+};
+typedef struct orte_fddp_trend_component_t orte_fddp_trend_component_t;
+
+ORTE_MODULE_DECLSPEC extern orte_fddp_trend_component_t mca_fddp_trend_component;
+extern orte_fddp_base_module_t orte_fddp_trend_module;
+
+
+END_C_DECLS
+
+#endif

orte/mca/fddp/trend/fddp_trend_component.c

@@ -0,0 +1,92 @@
+/*
+ * Copyright (c) 2004-2008 The Trustees of Indiana University and Indiana
+ *                         University Research and Technology
+ *                         Corporation.  All rights reserved.
+ * Copyright (c) 2004-2005 The University of Tennessee and The University
+ *                         of Tennessee Research Foundation.  All rights
+ *                         reserved.
+ * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart, 
+ *                         University of Stuttgart.  All rights reserved.
+ * Copyright (c) 2004-2005 The Regents of the University of California.
+ *                         All rights reserved.
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/base/base.h"
+#include "opal/mca/base/mca_base_param.h"
+#include "opal/class/opal_pointer_array.h"
+
+#include "orte/util/proc_info.h"
+#include "orte/util/show_help.h"
+
+#include "fddp_trend.h"
+
+/*
+ * Local functions
+ */
+
+static int orte_fddp_trend_open(void);
+static int orte_fddp_trend_close(void);
+static int orte_fddp_trend_query(mca_base_module_t **module, int *priority);
+
+orte_fddp_trend_component_t mca_fddp_trend_component = {
+    {
+        {
+            ORTE_FDDP_BASE_VERSION_1_0_0,
+            
+            "trend", /* MCA component name */
+            ORTE_MAJOR_VERSION,  /* MCA component major version */
+            ORTE_MINOR_VERSION,  /* MCA component minor version */
+            ORTE_RELEASE_VERSION,  /* MCA component release version */
+            orte_fddp_trend_open,  /* component open  */
+            orte_fddp_trend_close, /* component close */
+            orte_fddp_trend_query  /* component query */
+        },
+        {
+            /* The component is checkpoint ready */
+            MCA_BASE_METADATA_PARAM_CHECKPOINT
+        }
+    }
+};
+
+
+/**
+  * component open/close/init function
+  */
+static int orte_fddp_trend_open(void)
+{
+    mca_base_component_t *c = &mca_fddp_trend_component.super.base_version;
+
+    /* lookup parameters */
+    mca_base_param_reg_int(c, "window_size",
+                              "Size of sliding window to smooth data for trend [default: 1]",
+                              false, false, 80,  &mca_fddp_trend_component.window_size);
+
+    return ORTE_SUCCESS;
+}
+
+
+static int orte_fddp_trend_query(mca_base_module_t **module, int *priority)
+{    
+    *priority = 0;  /* select only if specified */
+    *module = (mca_base_module_t *)&orte_fddp_trend_module;
+    
+    return ORTE_SUCCESS;
+}
+
+/**
+ *  Close all subsystems.
+ */
+
+static int orte_fddp_trend_close(void)
+{    
+    return ORTE_SUCCESS;
+}
+

orte/mca/sensor/Makefile.am

@@ -0,0 +1,35 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+# main library setup
+noinst_LTLIBRARIES = libmca_sensor.la
+libmca_sensor_la_SOURCES =
+
+# header setup
+nobase_orte_HEADERS =
+
+# local files
+headers = sensor.h \
+          sensor_types.h
+
+libmca_sensor_la_SOURCES += $(headers)
+
+# Conditionally install the header files
+if WANT_INSTALL_HEADERS
+nobase_orte_HEADERS += $(headers)
+ortedir = $(includedir)/openmpi/orte/mca/sensor
+else
+ortedir = $(includedir)
+endif
+
+include base/Makefile.am
+
+distclean-local:
+	rm -f base/static-components.h

orte/mca/sensor/base/Makefile.am

@@ -0,0 +1,27 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+#
+# Additional copyrights may follow
+#
+# $HEADER$
+#
+
+headers += \
+        base/base.h
+
+libmca_sensor_la_SOURCES += \
+        base/sensor_base_open.c
+
+if !ORTE_DISABLE_FULL_SUPPORT
+
+headers += \
+        base/sensor_private.h
+
+libmca_sensor_la_SOURCES += \
+        base/sensor_base_close.c \
+        base/sensor_base_select.c \
+        base/sensor_base_scale.c
+
+endif

orte/mca/sensor/base/base.h

@@ -0,0 +1,62 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+/** @file:
+ */
+
+#ifndef MCA_SENSOR_BASE_H
+#define MCA_SENSOR_BASE_H
+
+/*
+ * includes
+ */
+#include "orte_config.h"
+
+#include "opal/class/opal_list.h"
+#include "opal/mca/mca.h"
+
+#include "orte/mca/sensor/sensor.h"
+
+
+/*
+ * Global functions for MCA overall collective open and close
+ */
+BEGIN_C_DECLS
+
+/*
+ * function definitions
+ */
+ORTE_DECLSPEC    int orte_sensor_base_open(void);
+ORTE_DECLSPEC    int orte_sensor_base_select(void);
+ORTE_DECLSPEC    int orte_sensor_base_close(void);
+
+/*
+ * globals that might be needed
+ */
+
+ORTE_DECLSPEC extern int orte_sensor_base_output;
+ORTE_DECLSPEC extern opal_list_t mca_sensor_base_components_available;
+ORTE_DECLSPEC extern opal_list_t orte_sensor_base_selected_modules;
+
+/* object definition */
+typedef struct {
+    opal_list_item_t super;
+    orte_sensor_base_component_t    *component;
+    orte_sensor_base_module_t       *module;
+} orte_sensor_base_selected_pair_t;
+OBJ_CLASS_DECLARATION(orte_sensor_base_selected_pair_t);
+
+#if !ORTE_DISABLE_FULL_SUPPORT
+
+/* no base functions to protect at this time */
+
+#endif /* ORTE_DISABLE_FULL_SUPPORT */
+
+END_C_DECLS
+#endif

orte/mca/sensor/base/sensor_base_close.c

@@ -0,0 +1,42 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include <stdio.h>
+
+#include "opal/mca/mca.h"
+#include "opal/mca/base/base.h"
+#include "orte/mca/sensor/base/base.h"
+
+
+int orte_sensor_base_close(void)
+{
+    orte_sensor_base_selected_pair_t *pair;
+    opal_list_item_t *item;
+    
+    /* destruct the list of modules so they each can finalize */
+    for (item = opal_list_get_first(&orte_sensor_base_selected_modules);
+         opal_list_get_end(&orte_sensor_base_selected_modules) != item;
+         item = opal_list_get_next(item)) {
+        pair = (orte_sensor_base_selected_pair_t*)item;
+        OBJ_DESTRUCT(pair);
+    }
+    
+    /* Close all remaining available components */
+    
+    mca_base_components_close(orte_sensor_base_output, 
+                              &mca_sensor_base_components_available, NULL);
+    
+    /* All done */
+    
+    return ORTE_SUCCESS;
+}

orte/mca/sensor/base/sensor_base_open.c

@@ -0,0 +1,125 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/mca.h"
+#include "opal/util/output.h"
+#include "opal/mca/base/base.h"
+#include "opal/mca/base/mca_base_param.h"
+
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "orte/mca/sensor/base/base.h"
+
+
+/*
+ * The following file was created by configure.  It contains extern
+ * statements and the definition of an array of pointers to each
+ * component's public mca_base_component_t struct.
+ */
+
+#include "orte/mca/sensor/base/static-components.h"
+
+/* object definition */
+static void construct(orte_sensor_base_selected_pair_t *obj)
+{
+    obj->component = NULL;
+    obj->module = NULL;
+}
+
+static void destruct(orte_sensor_base_selected_pair_t *obj)
+{
+    if (NULL != obj->module->finalize) {
+        obj->module->finalize();
+    }
+}
+
+OBJ_CLASS_INSTANCE(orte_sensor_base_selected_pair_t,
+                   opal_list_item_t,
+                   construct, destruct);
+
+/* base functions */
+static void start(void);
+static void stop(void);
+
+/*
+ * Global variables
+ */
+int orte_sensor_base_output = -1;
+orte_sensor_base_API_module_t orte_sensor = {
+    start,
+    stop
+};
+opal_list_t mca_sensor_base_components_available;
+opal_list_t orte_sensor_base_selected_modules;
+
+/**
+ * Function for finding and opening either all MCA components, or the one
+ * that was specifically requested via a MCA parameter.
+ */
+int orte_sensor_base_open(void)
+{
+    /* Debugging / verbose output.  Always have stream open, with
+       verbose set by the mca open system... */
+    orte_sensor_base_output = opal_output_open(NULL);
+    
+    /* construct the list of modules */
+    OBJ_CONSTRUCT(&orte_sensor_base_selected_modules, opal_list_t);
+    
+    /* Open up all available components */
+
+    if (ORTE_SUCCESS !=
+        mca_base_components_open("sensor", orte_sensor_base_output,
+                                 mca_sensor_base_static_components,
+                                 &mca_sensor_base_components_available, true)) {
+        return ORTE_ERROR;
+    }
+
+    /* All done */
+
+    return ORTE_SUCCESS;
+}
+
+static void start(void)
+{
+    orte_sensor_base_selected_pair_t *pair;
+    opal_list_item_t *item;
+    
+    for (item = opal_list_get_first(&orte_sensor_base_selected_modules);
+         opal_list_get_end(&orte_sensor_base_selected_modules) != item;
+         item = opal_list_get_next(item)) {
+        pair = (orte_sensor_base_selected_pair_t*)item;
+        if (NULL != pair->module->start) {
+            pair->module->start();
+        }
+    }
+    return;    
+}
+
+static void stop(void)
+{
+    orte_sensor_base_selected_pair_t *pair;
+    opal_list_item_t *item;
+    
+    for (item = opal_list_get_first(&orte_sensor_base_selected_modules);
+         opal_list_get_end(&orte_sensor_base_selected_modules) != item;
+         item = opal_list_get_next(item)) {
+        pair = (orte_sensor_base_selected_pair_t*)item;
+        if (NULL != pair->module->stop) {
+            pair->module->stop();
+        }
+    }
+    return;
+}

orte/mca/sensor/base/sensor_base_scale.c

@@ -0,0 +1,47 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/mca.h"
+#include "opal/util/output.h"
+#include "opal/mca/base/base.h"
+#include "opal/mca/base/mca_base_param.h"
+
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif
+
+#include "orte/mca/sensor/base/sensor_private.h"
+
+
+int orte_sensor_scale_data(orte_sensor_data_t *target, int num_values, float *data)
+{
+    int i;
+    
+    /* ensure we have enough data storage in the sensor data object */
+    if (NULL != target->data.bytes) {
+        /* clear out pre-existing data */
+        free(target->data.bytes);
+    }
+    /* allocate what we need */
+    target->data.bytes = (uint8_t*)malloc(num_values * sizeof(uint8_t));
+    memset(target->data.bytes, 0, num_values);
+    target->data.size = num_values;
+
+    /* convert the data */
+    for (i=0; i < num_values; i++) {
+        target->data.bytes[i] = UINT8_MAX * (data[i] - target->min) / (target->max - target->min);
+    }
+    
+    return ORTE_SUCCESS;
+}

orte/mca/sensor/base/sensor_base_select.c

@@ -0,0 +1,123 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/mca.h"
+#include "opal/mca/base/base.h"
+#include "opal/util/argv.h"
+
+#include "orte/mca/sensor/base/base.h"
+
+
+/**
+ * Function for weeding out sensor components that don't want to run.
+ *
+ * Call the init function on all available components to find out if
+ * they want to run.  Select all components that don't fail.  Failing
+ * components will be closed and unloaded.  The selected modules will
+ * be returned to the caller in a opal_list_t.
+ */
+int orte_sensor_base_select(void)
+{
+    mca_base_component_list_item_t *cli = NULL;
+    mca_base_component_t *component = NULL;
+    mca_base_module_t *module = NULL;
+    orte_sensor_base_module_t *nmodule;
+    opal_list_item_t *item;
+    int i, priority, ret;
+    char *include_list = NULL;
+    char **imodules = NULL;
+    orte_sensor_base_selected_pair_t *pair;
+    
+    /*
+     * Register the framework MCA param and look up include list
+     */
+    mca_base_param_reg_string_name("sensor", NULL,
+                                   "Comma-delimisted list of  sensor component to use (empty = all avail)",
+                                   false, false,
+                                   NULL, &include_list);
+    
+    /* if the list is empty, then we have nothing to do */
+    if (NULL == include_list) {
+        return ORTE_SUCCESS;
+    }
+    
+    /* separate the names of the sensors to be used */
+    imodules = opal_argv_split(include_list, ',');
+
+    /* Query all available components and ask if they have a module */
+    for (item = opal_list_get_first(&mca_sensor_base_components_available);
+         opal_list_get_end(&mca_sensor_base_components_available) != item;
+         item = opal_list_get_next(item)) {
+        cli = (mca_base_component_list_item_t *) item;
+        component = (mca_base_component_t *) cli->cli_component;
+        
+        /* If this component was not specified, skip it */
+        for (i = 0; NULL != imodules[i]; ++i) {
+            if (0 == strcmp(imodules[i], component->mca_component_name)) {
+                break;
+            }
+        }
+        if (NULL == imodules[i]) {
+            continue;
+        }
+        
+        /* If there's no query function, skip it */
+        if (NULL == component->mca_query_component) {
+            opal_output_verbose(5, orte_sensor_base_output,
+                                "mca:sensor:select: Skipping component [%s]. It does not implement a query function",
+                                component->mca_component_name );
+            continue;
+        }
+        
+        /* Query the component */
+        opal_output_verbose(5, orte_sensor_base_output,
+                            "mca:sensor:select: Querying component [%s]",
+                            component->mca_component_name);
+        ret = component->mca_query_component(&module, &priority);
+        
+        /* If no module was returned, then skip component */
+        if (ORTE_SUCCESS != ret || NULL == module) {
+            opal_output_verbose(5, orte_sensor_base_output,
+                                "mca:sensor:select: Skipping component [%s]. Query failed to return a module",
+                                component->mca_component_name );
+            continue;
+        }
+        
+        /* If we got a module, initialize it */
+        nmodule = (orte_sensor_base_module_t*) module;
+        if (NULL != nmodule->init) {
+            /* If the module doesn't want to be used, skip it */
+            if (ORTE_SUCCESS != (ret = nmodule->init()) ) {
+                if (NULL != nmodule->finalize) {
+                    nmodule->finalize();
+                }
+                continue;
+            }
+        }
+        
+        opal_output_verbose(5, orte_sensor_base_output,
+                            "mca:sensor:select: Adding component [%s] to active list",
+                            component->mca_component_name );
+
+        /* Make an item for the list */
+        pair = OBJ_NEW(orte_sensor_base_selected_pair_t);
+        pair->component = (orte_sensor_base_component_t*) component;
+        pair->module = nmodule;
+        
+        /* Add it to the list of operational sensors */
+        opal_list_append(&orte_sensor_base_selected_modules, &(pair->super));
+    }
+
+    return ORTE_SUCCESS;
+}

orte/mca/sensor/base/sensor_private.h

@@ -0,0 +1,40 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ */
+/** @file:
+ */
+
+#ifndef MCA_SENSOR_PRIVATE_H
+#define MCA_SENSOR_PRIVATE_H
+
+/*
+ * includes
+ */
+#include "orte_config.h"
+
+#include "opal/dss/dss_types.h"
+
+#include "orte/mca/sensor/sensor_types.h"
+
+
+/*
+ * Global functions for MCA overall collective open and close
+ */
+BEGIN_C_DECLS
+
+#if !ORTE_DISABLE_FULL_SUPPORT
+/*
+ * function definitions
+ */
+ORTE_DECLSPEC    int orte_sensor_scale_data(orte_sensor_data_t *target, int num_values, float *data);
+
+#endif /* ORTE_DISABLE_FULL_SUPPORT */
+
+END_C_DECLS
+#endif

orte/mca/sensor/pru/Makefile.am

@@ -0,0 +1,37 @@
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+dist_pkgdata_DATA = help-orte-sensor-pru.txt
+
+sources = \
+        sensor_pru.c \
+        sensor_pru.h \
+        sensor_pru_component.c
+
+# Make the output library in this directory, and name it either
+# mca_<type>_<name>.la (for DSO builds) or libmca_<type>_<name>.la
+# (for static builds).
+
+if OMPI_BUILD_sensor_pru_DSO
+component_noinst =
+component_install = mca_sensor_pru.la
+else
+component_noinst = libmca_sensor_pru.la
+component_install =
+endif
+
+mcacomponentdir = $(pkglibdir)
+mcacomponent_LTLIBRARIES = $(component_install)
+mca_sensor_pru_la_SOURCES = $(sources)
+mca_sensor_pru_la_LDFLAGS = -module -avoid-version
+
+noinst_LTLIBRARIES = $(component_noinst)
+libmca_sensor_pru_la_SOURCES =$(sources)
+libmca_sensor_pru_la_LDFLAGS = -module -avoid-version

orte/mca/sensor/pru/configure.params

@@ -0,0 +1,14 @@
+# -*- shell-script -*-
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved.
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+
+# Specific to this module
+
+PARAM_CONFIG_FILES="Makefile"

orte/mca/sensor/pru/help-orte-sensor-pru.txt

@@ -0,0 +1,52 @@
+# -*- text -*-
+#
+# Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved.
+#
+# $COPYRIGHT$
+# 
+# Additional copyrights may follow
+# 
+# $HEADER$
+#
+# This is the US/English general help file for the resilient mapper.
+#
+[orte-rmaps-resilient:alloc-error]
+There are not enough slots available in the system to satisfy the %d slots 
+that were requested by the application:
+  %s
+
+Either request fewer slots for your application, or make more slots available
+for use.
+[orte-rmaps-resilient:multi-apps-and-zero-np]
+RMAPS found multiple applications to be launched, with
+at least one that failed to specify the number of processes to execute.
+When specifying multiple applications, you must specify how many processes
+of each to launch via the -np argument.
+
+[orte-rmaps-resilient:per-node-and-too-many-procs]
+There are not enough nodes in your allocation to satisfy your request to launch
+%d processes on a per-node basis - only %d nodes were available.
+
+Either request fewer processes, or obtain a larger allocation.
+[orte-rmaps-resilient:n-per-node-and-too-many-procs]
+There are not enough nodes in your allocation to satisfy your request to launch
+%d processes on a %d per-node basis - only %d nodes with a total of %d slots were available.
+
+Either request fewer processes, or obtain a larger allocation.
+[orte-rmaps-resilient:n-per-node-and-not-enough-slots]
+There are not enough slots on the nodes in your allocation to satisfy your request to launch on a %d process-per-node basis - only %d slots/node were available.
+
+Either request fewer processes/node, or obtain a larger allocation.
+
+[orte-rmaps-resilient:no-np-and-user-map]
+You have specified a rank-to-node/slot mapping, but failed to provide
+the number of processes to be executed. For some reason, this information
+could not be obtained from the mapping you provided, so we cannot continue
+with executing the specified application.
+#
+[orte-rmaps-resilient:file-not-found]
+The specified file that describes the fault groups for this system:
+
+FILE: %s
+
+was not found. Please verify the file name and location.

orte/mca/sensor/pru/sensor_pru.c

@@ -0,0 +1,56 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+#include "orte/types.h"
+
+#include <errno.h>
+#ifdef HAVE_UNISTD_H
+#include <unistd.h>
+#endif  /* HAVE_UNISTD_H */
+#ifdef HAVE_STRING_H
+#include <string.h>
+#endif  /* HAVE_STRING_H */
+#include <stdio.h>
+
+#include "opal/mca/base/mca_base_param.h"
+#include "opal/util/argv.h"
+
+#include "orte/util/show_help.h"
+#include "orte/mca/errmgr/errmgr.h"
+
+#include "sensor_pru.h"
+
+/* declare the functions */
+static int start(void);
+static int stop(void);
+
+/* instantiate the module */
+orte_sensor_base_module_t orte_sensor_pru_module = {
+    start,
+    stop
+};
+
+
+/*
+ * Start monitoring of local processes
+ */
+static int start(void)
+{
+    return ORTE_SUCCESS;
+}
+
+
+static int stop(void)
+{
+    return ORTE_SUCCESS;
+}
+

orte/mca/sensor/pru/sensor_pru.h

@@ -0,0 +1,36 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+/**
+ * @file
+ *
+ * Process Resource Utilization sensor 
+ */
+#ifndef ORTE_SENSOR_PRU_H
+#define ORTE_SENSOR_PRU_H
+
+#include "orte_config.h"
+
+#include "orte/mca/sensor/sensor.h"
+
+BEGIN_C_DECLS
+
+struct orte_sensor_pru_component_t {
+    orte_sensor_base_component_t super;
+    int sample_rate;
+};
+typedef struct orte_sensor_pru_component_t orte_sensor_pru_component_t;
+
+ORTE_MODULE_DECLSPEC extern orte_sensor_pru_component_t mca_sensor_pru_component;
+extern orte_sensor_base_module_t orte_sensor_pru_module;
+
+
+END_C_DECLS
+
+#endif

orte/mca/sensor/pru/sensor_pru_component.c

@@ -0,0 +1,91 @@
+/*
+ * Copyright (c) 2004-2008 The Trustees of Indiana University and Indiana
+ *                         University Research and Technology
+ *                         Corporation.  All rights reserved.
+ * Copyright (c) 2004-2005 The University of Tennessee and The University
+ *                         of Tennessee Research Foundation.  All rights
+ *                         reserved.
+ * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart, 
+ *                         University of Stuttgart.  All rights reserved.
+ * Copyright (c) 2004-2005 The Regents of the University of California.
+ *                         All rights reserved.
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#include "opal/mca/base/base.h"
+#include "opal/mca/base/mca_base_param.h"
+#include "opal/class/opal_pointer_array.h"
+
+#include "orte/util/proc_info.h"
+#include "orte/util/show_help.h"
+
+#include "sensor_pru.h"
+
+/*
+ * Local functions
+ */
+
+static int orte_sensor_pru_open(void);
+static int orte_sensor_pru_close(void);
+static int orte_sensor_pru_query(mca_base_module_t **module, int *priority);
+
+orte_sensor_pru_component_t mca_sensor_pru_component = {
+    {
+        {
+            ORTE_SENSOR_BASE_VERSION_1_0_0,
+            
+            "pru", /* MCA component name */
+            ORTE_MAJOR_VERSION,  /* MCA component major version */
+            ORTE_MINOR_VERSION,  /* MCA component minor version */
+            ORTE_RELEASE_VERSION,  /* MCA component release version */
+            orte_sensor_pru_open,  /* component open  */
+            orte_sensor_pru_close, /* component close */
+            orte_sensor_pru_query  /* component query */
+        },
+        {
+            /* The component is checkpoint ready */
+            MCA_BASE_METADATA_PARAM_CHECKPOINT
+        }
+    }
+};
+
+
+/**
+  * component open/close/init function
+  */
+static int orte_sensor_pru_open(void)
+{
+    mca_base_component_t *c = &mca_sensor_pru_component.super.base_version;
+
+    /* lookup parameters */
+    mca_base_param_reg_int(c, "sample_rate",
+                              "Sample rate in seconds (default=10)",
+                              false, false, 10,  &mca_sensor_pru_component.sample_rate);
+    
+    return ORTE_SUCCESS;
+}
+
+
+static int orte_sensor_pru_query(mca_base_module_t **module, int *priority)
+{    
+    *priority = 0;  /* select only if specified */
+    *module = (mca_base_module_t *)&orte_sensor_pru_module;
+    
+    return ORTE_SUCCESS;
+}
+
+/**
+ *  Close all subsystems.
+ */
+
+static int orte_sensor_pru_close(void)
+{
+    return ORTE_SUCCESS;
+}

orte/mca/sensor/sensor.h

@@ -0,0 +1,96 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ *
+ * @file:
+ *
+ */
+
+#ifndef MCA_SENSOR_H
+#define MCA_SENSOR_H
+
+/*
+ * includes
+ */
+
+#include "orte_config.h"
+
+#include "opal/mca/mca.h"
+
+BEGIN_C_DECLS
+
+/*
+ * Component functions - all MUST be provided!
+ */
+
+/* initialize the selected module */
+typedef int (*orte_sensor_base_module_init_fn_t)(void);
+    
+/* finalize the selected module */
+typedef void (*orte_sensor_base_module_finalize_fn_t)(void);
+
+/* start collecting data */
+typedef void (*orte_sensor_base_module_start_fn_t)(void);
+
+/* stop collecting data */
+typedef void (*orte_sensor_base_module_stop_fn_t)(void);
+
+/* API module */
+/*
+ * Ver 1.0
+ */
+struct orte_sensor_base_API_module_1_0_0_t {
+    orte_sensor_base_module_start_fn_t      start;
+    orte_sensor_base_module_stop_fn_t       stop;
+};
+
+typedef struct orte_sensor_base_API_module_1_0_0_t orte_sensor_base_API_module_1_0_0_t;
+typedef orte_sensor_base_API_module_1_0_0_t orte_sensor_base_API_module_t;
+
+
+/*
+ * Component modules Ver 1.0
+ */
+struct orte_sensor_base_module_1_0_0_t {
+    orte_sensor_base_module_init_fn_t       init;
+    orte_sensor_base_module_finalize_fn_t   finalize;
+    orte_sensor_base_module_start_fn_t      start;
+    orte_sensor_base_module_stop_fn_t       stop;
+};
+
+typedef struct orte_sensor_base_module_1_0_0_t orte_sensor_base_module_1_0_0_t;
+typedef orte_sensor_base_module_1_0_0_t orte_sensor_base_module_t;
+
+/*
+ * the standard component data structure
+ */
+struct orte_sensor_base_component_1_0_0_t {
+    mca_base_component_t base_version;
+    mca_base_component_data_t base_data;
+};
+typedef struct orte_sensor_base_component_1_0_0_t orte_sensor_base_component_1_0_0_t;
+typedef orte_sensor_base_component_1_0_0_t orte_sensor_base_component_t;
+
+
+
+/*
+ * Macro for use in components that are of type sensor v1.0.0
+ */
+#define ORTE_SENSOR_BASE_VERSION_1_0_0 \
+  /* sensor v1.0 is chained to MCA v2.0 */ \
+  MCA_BASE_VERSION_2_0_0, \
+  /* sensor v1.0 */ \
+  "sensor", 1, 0, 0
+
+/* Global structure for accessing sensor functions
+ */
+ORTE_DECLSPEC extern orte_sensor_base_API_module_t orte_sensor;  /* holds API function pointers */
+
+END_C_DECLS
+
+#endif /* MCA_SENSOR_H */

orte/mca/sensor/sensor_types.h

@@ -0,0 +1,54 @@
+/*
+ * Copyright (c) 2009      Cisco Systems, Inc.  All rights reserved. 
+ *
+ * $COPYRIGHT$
+ * 
+ * Additional copyrights may follow
+ * 
+ * $HEADER$
+ */
+/** @file:
+ */
+
+#ifndef ORTE_MCA_SENSOR_TYPES_H
+#define ORTE_MCA_SENSOR_TYPES_H
+
+#include "orte_config.h"
+#include "orte/constants.h"
+
+#ifdef HAVE_SYS_TIME_H
+#include <sys/time.h>
+#endif  /* HAVE_SYS_TIME_H */
+
+#include "opal/dss/dss_types.h"
+
+/*
+ * General SENSOR types - instanced in runtime/orte_globals.c
+ */
+
+BEGIN_C_DECLS
+
+enum {
+    ORTE_SENSOR_SCALE_LINEAR,
+    ORTE_SENSOR_SCALE_LOG,
+    ORTE_SENSOR_SCALE_SIGMOID
+};
+
+/*
+ * Structure for passing data from sensors
+ */
+typedef struct {
+    opal_object_t super;
+    char *sensor;
+    struct timeval timestamp;
+    int scaling_law;
+    float min;
+    float max;
+    float gain;
+    opal_byte_object_t data;
+} orte_sensor_data_t;
+ORTE_DECLSPEC OBJ_CLASS_DECLARATION(orte_sensor_data_t);
+
+END_C_DECLS
+
+#endif

orte/runtime/orte_globals.c

@@ -35,6 +35,7 @@
 #include "opal/dss/dss.h"
 
 #include "orte/mca/errmgr/errmgr.h"
+#include "orte/mca/sensor/sensor_types.h"
 #include "orte/util/proc_info.h"
 #include "orte/util/name_fns.h"
 
@@ -969,4 +970,27 @@ OBJ_CLASS_INSTANCE(orte_regex_node_t,
                    orte_regex_node_construct,
                    orte_regex_node_destruct);
 
+static void orte_sensor_data_construct(orte_sensor_data_t *ptr)
+{
+    ptr->sensor = NULL;
+    ptr->scaling_law = ORTE_SENSOR_SCALE_LINEAR;
+    ptr->min = 0.0;
+    ptr->max = 100.0;
+    ptr->gain = 1.0;
+    ptr->data.size = 0;
+    ptr->data.bytes = NULL;
+}
+static void orte_sensor_data_destruct(orte_sensor_data_t *ptr)
+{
+    if (NULL != ptr->sensor) {
+        free(ptr->sensor);
+    }
+    if (NULL != ptr->data.bytes) {
+        free(ptr->data.bytes);
+    }
+}
+OBJ_CLASS_INSTANCE(orte_sensor_data_t,
+                   opal_object_t,
+                   orte_sensor_data_construct,
+                   orte_sensor_data_destruct);
 #endif