Knowledge Driven Dynamic Frequency Scaling
Abstract
This paper proposes a novel, knowledge driven dynamic frequency scaling approach for dynamic power management. It uses a neural network model for learning the behavior of various applications and then uses this knowledge for predicting the optimal processor frequency at any instance of time. The experimental results show the elegance of the approach. The approach may be suitably modified for application in multicore processors and SoCs(System on Chips).
Details
| Title: | Knowledge Driven Dynamic Frequency Scaling |
| Subjects: | Engineering |
| More Details: | View PDF |
| Report Article: | Report |
Submission History
From:
Anirban Lahiri
[View Profile]
Date of Publication:
July 15, 2025, 8:21 a.m. UTC