} s_probabilist_event_generator_t;
typedef struct tmgr_trace {
- enum e_trace_type type;
- union {
- struct {
- xbt_dynar_t event_list;
- } s_list;
- struct {
- probabilist_event_generator_t event_generator[2];
- int is_state_trace;
- int next_event;
- } s_probabilist;
- };
+ xbt_dynar_t event_list;
} s_tmgr_trace_t;
/* Iterator within a trace */
#ifdef __cplusplus
namespace simgrid {
+/** @brief Modeling of the resource variations, such as those due to an external load
+ *
+ * There is 3 main concepts in this module:
+ * - #trace: a set of dated values, ie a list of pair <timestamp, value>
+ * - #trace_iterator: links a given trace to a given simgrid resource. A Cpu for example has 2 iterators: state (ie, is it ON/OFF) and speed, while a link has 3 iterators: state, bandwidth and latency.
+ * - #future_evt_set: makes it easy to find the next occuring event of all traces
+ */
namespace trace_mgr {
-/* Future Event Set (collection of iterators over the traces)
+/** @brief A trace_iterator links a trace to a resource */
+XBT_PUBLIC_CLASS trace_iterator {
+
+};
+
+/** @brief A trace is a set of timed values, encoding the value that a variable takes at what time *
+ *
+ * It is useful to model dynamic platforms, where an external load that makes the resource availability change over time.
+ * To model that, you have to set several traces per resource: one for the on/off state and one for each numerical value (computational speed, bandwidt and latency).
+ */
+XBT_PUBLIC_CLASS trace {
+public:
+ /** Creates an empty trace */
+ trace() {}
+ virtual ~trace() {}
+};
+
+/** @brief Future Event Set (collection of iterators over the traces)
* That's useful to quickly know which is the next occurring event in a set of traces. */
XBT_PUBLIC_CLASS future_evt_set {
public: