The performance of a GPU is often measured in the number of polygons per second being...

 

Transcribed Image Text:

The performance of a GPU is often measured in the number of polygons per second being rendered. Consider a GPU with 2 processing cores, each capable of rendering a polygon in parallel. A video game requires on average 100 polygons per millisecond to be rendered. Each core can render a polygon in 15 microseconds. Polygons are placed in a single queue shared for the two GPU cores. Answer the following: a) What queuing system would you use to model this system? State any assumptions you need to make for the model to hold. b) If each polygon can be described with 3 pairs of 64-bit floating point numbers, how much video memory (in bytes) is being used on average? Consider that a polygon being rendered still occupies space in memory. c) What is the probability that the GPU will be idle? d) What is the average amount of time it takes for a polygon to be rendered? The performance of a GPU is often measured in the number of polygons per second being rendered. Consider a GPU with 2 processing cores, each capable of rendering a polygon in parallel. A video game requires on average 100 polygons per millisecond to be rendered. Each core can render a polygon in 15 microseconds. Polygons are placed in a single queue shared for the two GPU cores. Answer the following: a) What queuing system would you use to model this system? State any assumptions you need to make for the model to hold. b) If each polygon can be described with 3 pairs of 64-bit floating point numbers, how much video memory (in bytes) is being used on average? Consider that a polygon being rendered still occupies space in memory. c) What is the probability that the GPU will be idle? d) What is the average amount of time it takes for a polygon to be rendered?