APS View

Note:

QNX OS 8.0 does not support adaptive partitioning; if your target is based on 8.0, then the IDE won't receive any APS data and this pane isn't meaningful. For targets based on an earlier OS version, this pane is meaningful only when the target uses adaptive partitioning.

For more information about adaptive partitioning, see the following content in the user documentation for the OS version that your target is based on:

the Adaptive Partitioning chapter of the System Architecture guide
the Adaptive Partitioning User's Guide
the aps entry in the Utilities Reference

To configure APS for a connected target, you must use the Target Navigator. The APS View shows APS-related statistics and settings in multiple panels within its top pane; the subsections that follow describe these panels. The bottom pane displays graphs of the partition budgets and CPU usage.

Screenshot of APS View showing Partitions panel, which reveals parent partition and budget for all partitions, and Partition Budgets piechart, which illustrates CPU resource assignments to all partitions

APS System Information

This panel displays the following APS state fields:

current number of partitions
partition, process, and thread IDs where the last bankruptcy (exhaustion of critical budget) occurred
average number of computational cycles executed per millisecond
averaging windows, which are time intervals used by the scheduler to calculate CPU usage
APS security flags in effect, which determine who can add partitions and modify their attributes
relative and overall times that the system was idle for longer than the second and third averaging windows (by default, these are 1 and 10 seconds)

Note:

When you add or modify partitions using the SchedCtl() C library function or the aps utility, the changes are reflected in this and all other panels when the view is refreshed.

Partitions

In this panel, the Partition Information table shows these details for each partition:

partition name
parent partition
partition ID
budget, which is the percentage of CPU time guaranteed for the partition
critical budget, which is the time that critical threads can run over budget

On the right, the Partition Child Processes and Threads area lists the processes and threads assigned to each partition. The list is interactive, allowing you to move processes and threads between partitions by dragging and dropping. Moving one of these items might cause others to be moved as well.

Partition Statistics

This panel displays CPU usage statistics related to the averaging windows. The table shows the relative CPU times during which threads in each partition ran for longer than the second and third windows. It also shows the total CPU times that all or just critical thread runtimes exceeded those two windows.

By default, the second and third averaging windows are 1 and 10 seconds, but you can change them by setting a new first window, through SchedCtl() or aps. The second window is set to 10 times this duration and the third window to 100 times.

APS Bankruptcy

In this panel, the Bankruptcy Information table lists the IDs of each partition's notification process and thread (which handle the event emitted when the partition exhausts its critical budget and thus, becomes bankrupt), and the IDs of the process and thread running when the last bankruptcy occurred.

On the right, the Bankruptcy Flags area lists the system's bankruptcy settings, which determine how it reacts to a bankruptcy. You can change the bankruptcy settings through SchedCtl() or aps.

Partition and CPU usage graphs

The APS View contains the following graphs in multiple tabs:

Partition Budgets: This piechart shows the partition budgets, which are the percentages of CPU time assigned to the partitions.
CPU Usage: This line chart shows recent CPU usage for each partition.
Critical Time Usage: This line chart shows critical time usage for each partition, which refers to the time (in milliseconds) that critical threads within a partition ran over budget.