| Prev | Next | Start of Chapter | Next Chapter | Contents | Glossary | Index | Comments | (9 out of 9)

Types of G2-Meters

The following table lists all G2-meters. A brief description of each type of meter follows the table. Some general considerations about memory meters appear in the previous section. All meters can have lagged values except where otherwise noted.

Memory Meters Time Meters
memory-size
clock-tick-length
memory-usage
maximum-clock-tick-length
memory-available
percent-run-time
region-1-memory-size
simulator-time-lag
region-1-memory-usage
priority-1-scheduler-time-lag
region-1-memory-available
priority-2-scheduler-time-lag
region-2-memory-size
priority-3-scheduler-time-lag
region-2-memory-usage
priority-4-scheduler-time-lag
region-2-memory-available
priority-5-scheduler-time-lag
region-3-memory-size
priority-6-scheduler-time-lag
region-3-memory-usage
priority-7-scheduler-time-lag
region-3-memory-available
priority-8-scheduler-time-lag

priority-9-scheduler-time-lag

priority-10-scheduler-time-lag

The G2 Simulator is a superseded capability. For more information, see Appendix F, Superseded Practices.

memory-size

Measures the total memory allocated to G2 by the operating system for holding data. The figure includes both used and available memory. The sum of memory-usage and memory-available should equal memory-size.

memory-usage

Measures the total amount of memory that G2 currently uses. As G2 creates more items, schedules more tasks, and the like, memory usage increases.

memory-available

Measures the total amount of memory currently allocated by the operating system but not used by G2.

region-n-memory-size

Each meter measures the memory in the G2 region specified by n. The figure includes both used and available memory. For each region, the sum of memory available and memory usage should equal memory size.

region-n-memory-usage

Each meter measures the amount of memory that G2 currently uses in the G2 region specified by n.

region-n-memory-available

Each meter measures the total amount of memory currently available to G2 but not used by it in the G2 region specified by n.

clock-tick-length

Computes how many seconds a G2 clock tick lasts. If the scheduler mode is real time, this value varies closely around 1 second. In other modes, clock-tick-length may vary more widely.

maximum-clock-tick-length

Holds the duration in seconds of the longest clock tick that G2 has experienced since the knowledge base started running. If the scheduler mode is not real time, the maximum clock tick length can be large, for example if you pause the knowledge base for a long time. Maximum-clock-tick-length is not a lagged value.

percent-run-time

Computes how much processing time G2 is using, as a percent of the processing time available for it to use. The value of the meter is recalculated at each clock tick by computing:

The time that G2 sleeps even though it could be processing acts as a reserve of processing power on which it could draw if the demand on it increased. The percent-run-time meter measures the size of that reserve: the lower the value, the more reserve exists.

In general, the percent-run-time should not be more than 80% unless you have used priorities to be sure that high-priority tasks will get done even if demand exceeds available time, causing G2 to lag.

simulator-time-lag

Computes how many seconds behind the current system time the G2 simulator is. If the value is positive, this means that the simulator is behind schedule; if it is negative, the simulator is ahead. This value cannot be positive if the scheduler mode is simulated time. Simulator-time-lag is not a lagged value.

The G2 Simulator is a superseded capability. For more information, see Appendix F, Superseded Practices.

priority-n-scheduler-time-lag

These meters compute how many seconds behind current system time the scheduler is for a given priority. There are ten such meters, each representing the queue of tasks of a particular priority, within the current task queue. Tasks of priority one are first in the current task queue, then tasks of priority two, and so on.

If the value of one of these meters is positive, the scheduler is behind schedule on that priority's queue. For example, if the value of the priority-1-scheduler-time-lag meter is 0 and the value of the priority-2-scheduler-time-lag meter is 3, then all tasks of priority one are being completed, but task of priority two (and probably the lower priorities, as well) are behind by three clock ticks.

The value of any of these meters will always be 0 when the scheduler mode is either simulated time or as fast as possible. Scheduler-time-lag is not a lagged value.

| Prev | Next | Start of Chapter | Next Chapter | Contents | Glossary | Index | Comments | (9 out of 9)

Copyright © 1997 Gensym Corporation, Inc.