{ Views Help - Utilization View }
{ The Utilization View summarizes the Space-Time View }
{ at each time instant, showing the number of tasks }
{ Computing, in Overhead, or Waiting for a message at }
{ any given time. }
{}
{ This information is represented by three colored }
{ rectangles, stacked vertically at each time instant, }
{ with Computing on the bottom, Overhead in the middle, }
{ and Waiting on top.  The default colors are green for }
{ Computing, yellow for Overhead, and red for Waiting. }
{}
{ The overall utilization for tasks is seen by comparing }
{ the relative height of the rectangles over time, such }
{ that in cases of good utilization the Computing areas }
{ are prominent, and for poor utilization the Overhead }
{ and Waiting areas dominate. }
{}
{ The Utilization and Space-Time Views share the same }
{ horizontal time scale, and entering either view area }
{ with the mouse pointer creates a blue time line in }
{ both views.  This lines identifies the same time }
{ instant in each view's display, to help correlate }
{ the two representations. }
{}
{ Zooming }
{ ------- }
{}
{ The Utilization View can be "zoomed" in or out to view }
{ different perspectives on the same trace data.  To zoom }
{ in on a particular area of the Utilization view, click }
{ and hold on the main canvas with the Middle mouse button, }
{ and sweep across the area that is to be blown up.  When }
{ the mouse button is released, the swept area will be }
{ enlarged horizontally to fill the canvas area. }
{}
{ To "unzoom" back to the previous zoom level, or out to some }
{ lower magnification level if no zoom had been previously }
{ applied, simply press the Right mouse button. }
{}
{ To summarize: }
{}
{	   Mouse Middle        Mouse Right }
{}
{	     Zoom IN             Zoom Out }
{}
{ Like the time lines above, the Space-Time View is }
{ automagically zoomed in or out to match the Utilization }
{ View's time scale. }
{}
{ It should be noted that the Utilization View has been }
{ significantly optimized in XPVM 1.1.  A task's status }
{ is not added to the utilization task counts until that }
{ task's state is known for a particular time range.  This }
{ results in a "trailing edge" as tasks gradually fill in }
{ their events from left to right.  The overall result is }
{ a more correct depiction of the true task utilization }
{ that is significantly more efficient, by several orders }
{ of magnitude. }
{}