English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

As the functions of mobile devices are diversified, the power consumption of System on Chip (SoC) is increasing rapidly. Among the various types of SoC, Application Processor (AP) controls all functions of mobile device and is the most power-consuming SoC. In the AP manufacturing process, evaluating the characteristics of semiconductor in order to apply the optimal voltage that allows AP to have the lowest power consumption has a great influence on the AP quality. However, it is practically impossible to measure the lowest voltage for individual APs due to considerations such as yield and evaluation time during the manufacturing process. Therefore, voltage binning technique that applies the same voltage by dividing APs into groups according to semiconductor characteristics has traditionally been used, but this can cause a large difference from the lowest voltage that individual APs can actually reach. Therefore, a new voltage prediction model is needed to reduce the over-supplied voltage and accurately estimate the lowest voltage for individual APs. In this paper, we propose a minimum voltage prediction model based on Deep Neural Network (DNN) to evaluate the lowest voltage of APs. We use the leakage current, the ring oscillator and the operation temperature, which are characteristic values of each AP, as inputs to create our prediction model that considers the relationship between each element and the minimum voltage. By using our model, we can find and apply the optimal voltage to each AP so that it operates at lower voltage than the traditional method. Ultimately, we have demonstrated that using the new model can improve the applied voltage by up to 6.76%.

Minimum Voltage Prediction Model for Application Processor Based on Deep Neural Network in Manufacturing Process