Browsing by Author "Chin, Ting-Wu"

Now showing 1 - 2 of 2

Domain-Specific Approximation for Object Detection
(2018-01-01)
Chin, Ting-Wu
;
Yu, Chia-Lin
;
Halpern, Matthew
;
Genc, Hasan
;
Tsao, Shiao-Li
There is growing interest in object detection in advanced driver assistance systems and autonomous robots and vehicles. To enable such innovative systems, we need faster object detection. In this article, we investigate the trade-off between accuracy and speed with domain-specific approximations (DSAs) for two state-of-the-art deep learning-based object detection meta-architectures. We study the effectiveness of applying approximation both statically and dynamically to understand their potential and applicability. By conducting experiments on the ImageNet VID dataset, we show that DSA has great potential to improve the speed of the system without deteriorating the accuracy of object detectors. To this end, we present our insights on harvesting DSA and devise a proof-of-concept runtime, AutoFocus, that exploits dynamic DSA.
Improving the Accuracy of the Leakage Power Estimation of Embedded CPUs
(2017-01-01)
Chin, Ting-Wu
;
Tsao, Shiao-Li
;
Hung, Kuo-Wei
;
Huang, Pei-Shu
;
資訊工程學系
Previous studies have used on-chip thermal sensors (diodes) to estimate the leakage power of a CPU. However, an embedded CPU equips only a few thermal sensors and may suffer from considerable spatial temperature variances across the CPU core, and leakage power estimation based on insufficient temperature information introduces errors. According to our experiments, the conventional leakage power models may have up to 22.9% estimation error for a 70-nm embedded CPU. In this study, we first evaluated the accuracy of leakage power estimates based on thermal sensors on different locations of a CPU and suggested locations that can reduce the error to 0.9%. Then, we proposed temperature-referred and counter-tracked estimation (TRACE) that relies on temperature sensors and hardware activity counters to estimate leakage power. The simulation results demonstrated that employing TRACE could reduce the error to 3.4%. Experiments were also conducted on a real platform to verify our findings.