Abstract
Many application scenarios of edge visual inference, e.g., robotics or environmental monitoring, eventually require long periods of continuous operation. In such periods, the processor temperature plays a critical role to keep a prescribed frame rate. Particularly, the heavy computational load of convolutional neural networks (CNNs) may lead to thermal throttling and hence performance degradation in few seconds. In this paper, we report and analyze the long-term performance of 80 different cases resulting from running five CNN models on four software frameworks and two operating systems without and with active cooling. This comprehensive study was conducted on a low-cost edge platform, namely Raspberry Pi 4B (RPi4B), under stable indoor conditions. The results show that hysteresis-based active cooling prevented thermal throttling in all cases, thereby improving the throughput up to approximately 90% versus no cooling. Interestingly, the range of fan usage during active cooling varied from 33% to 65%. Given the impact of the fan on the power consumption of the system as a whole, these results stress the importance of a suitable selection of CNN model and software components. To assess the performance in outdoor applications, we integrated an external temperature sensor with the RPi4B and conducted a set of experiments with no active cooling in a wide interval of ambient temperature, ranging from 22 °C to 36 °C. Variations up to 27.7% were measured with respect to the maximum throughput achieved in that interval. This demonstrates that ambient temperature is a critical parameter in case active cooling cannot be applied.
Highlights
Deep learning (DL) [1] and its particular embodiment in the form of convolutional neural networks (CNNs) have become the de-facto approach to address many computer vision tasks, e.g., image recognition, object detection, or segmentation
According to the results presented in the previous section, thermal throttling can be prevented through active cooling
We conducted a number of outdoor tests to evaluate the impact of ambient temperature sweeping a wide variation interval while performing continuous CNN processing on the Raspberry Pi 4B (RPi4B) with no active cooling
Summary
New training and processing techniques together with the availability of large datasets and high computational power are the main underlying factors supporting the distinctive feature of CNNs versus classical algorithms, i.e., their high accuracy. Reducing the computational load of CNNs through techniques such as network pruning, data quantization, and network compression [9,10,11]. Despite these efforts, the implementation of CNNs still constitutes a major challenge for edge visual inference. The implementation of CNNs still constitutes a major challenge for edge visual inference Application scenarios such as robotics [12], environmental monitoring [13], Electronics 2020, 9, 2106; doi:10.3390/electronics9122106 www.mdpi.com/journal/electronics. DL-based processing pipelines deplete the scarce resources available in such devices, significantly affecting the timely completion of other tasks
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call