Other examples might come from robotics. For example, needing to compute a robot's answer to a human's question within the time expected for a response, or compute its best next motion before it falls off a cliff. Game-playing algorithms like chess would also fall in this category.
This comment was marked helpful 0 times.
stephyeung
TC scaling is useful when you're able to do things in an amount of time that matters to you.
Most important time thresholds: interactive rates, over lunch break (i.e. come back after lunch and your results are ready for you), and overnight :)
This comment was marked helpful 0 times.
LeeK
In the assignment 2 example, the time constraint is 30 FPS since anything less than that creates a visible stutter on the screen. So, for this type of scaling, we want to know how well more processors can do more work and render more complex scenes given the time constraint.
This comment was marked helpful 0 times.
toastifer
For assignment 2, we don't want use our faster GPU to render more frames because the increase in frames-per-second is negligible to our eyes. By rendering more complex scenes instead, we get a more perceivable use for our faster GPU.
Other examples might come from robotics. For example, needing to compute a robot's answer to a human's question within the time expected for a response, or compute its best next motion before it falls off a cliff. Game-playing algorithms like chess would also fall in this category.
This comment was marked helpful 0 times.
TC scaling is useful when you're able to do things in an amount of time that matters to you. Most important time thresholds: interactive rates, over lunch break (i.e. come back after lunch and your results are ready for you), and overnight :)
This comment was marked helpful 0 times.
In the assignment 2 example, the time constraint is 30 FPS since anything less than that creates a visible stutter on the screen. So, for this type of scaling, we want to know how well more processors can do more work and render more complex scenes given the time constraint.
This comment was marked helpful 0 times.
For assignment 2, we don't want use our faster GPU to render more frames because the increase in frames-per-second is negligible to our eyes. By rendering more complex scenes instead, we get a more perceivable use for our faster GPU.
This comment was marked helpful 0 times.