This workload explores the scaling of the number of scripted objects with periodic updates in a scene and their affect on platform metrics including CPU and network utilization.
The scene is a blank flat square region with an object generator in the middle. Each time the command is given, the generator creates 100 objects. Each of those objects changes its rotation and color at random intervals between 1-5 seconds.
The TestClient connects 10 avatars and then repeats the following steps until 1000 updating objects have been created.
more updates causing the object generator to create 100 objectsNote: The object generator will not allow itself to be activated more often than every 10 seconds. The effect is that it responds to the first TestClient avatar it hears and ignores the rest.
The spikes in CPU during the object creation phase is apparent. Both the TX bandwidth and CPU usage increase in steps as each set of 100 objects is created. At 1000 objects steady state:
Note: The bandwith usage is, as expected, exactly 10 times greater than the 1 avatar version of this workload, Objects01.