Traditional
graphics architecture. |
||
•Optimized for OpenGL-style fast texturing |
||
•Very deep pipeline (around 200 stages) |
||
•Data through all stages |
||
•All stages always clocking, utilizing power
regardless of whether or not they are actually doing useful work. |
||
•Typically these traditional pipelines use
around 750mW per 100 mega-pixels per second of graphics work |
||
•Very power hungry |
||
•Very fast because of dedicated units but power
intensive |
||
•Not well-suited for low-powered devices |