| Paper: | TA-P3.8 |
| Session: | Image and Video Modeling II |
| Time: | Tuesday, September 18, 09:50 - 12:30 |
| Presentation: |
Poster
|
| Title: |
GRACEFUL QUALITY DEGRADATION FOR VIDEO DECODING SYSTEMS THROUGH PRIORITY SCHEDULING AND PROCESSOR POWER ADAPTATION |
| Authors: |
Brian Foo; University of California, Los Angeles | | |
| | Mihaela van der Schaar; University of California, Los Angeles | | |
| Abstract: |
Voltage/frequency configurable processors can provide significant energy savings in video decoding systems due to their ability to dynamically adapting the frequency and voltage according to time-varying workloads. In this paper, we propose a joint voltage scaling and priority scheduling algorithm that decodes jobs in order of their importance (quality impact), such that by setting the processor to various power levels and decoding only the most important jobs, different quality and energy tradeoffs can be achieved. We demonstrate that our algorithm performs well in practical decoding scenarios, where reducing the power to 25% of the original power can lead to quality degradations of less than 1.0 dB PSNR. |
