Xianglan Chen

Definitions of predictability for Cyber Physical Systems

With the recent proliferation of different types of Cyber Physical Systems (CPS), it is critically important to investigate the predictability of such systems. Along with functional correctness of the components, these systems must also ensure that timing and delay constraints of components are properly for the entire system to behave in a predictable manner in presence of various kinds of uncertainties. While the functional correctness of the CPS components has been investigated in the past, very little is available about the timing issues. The objective of this paper is to conduct an investigation of key issues involved to ensure the predictability of the system, introduce rigorous definitions of performance parameters, and propose metrics for their evaluation and analyze their suitability to be used in the presence of uncertainties in which CPS operate. The results are expected to provide greater insight into the time critical behavior of CPS components.

MuDBN: An Energy-Efficient and High-Performance Multi-FPGA Accelerator for Deep Belief Networks

With the increasing size of neural networks, state-of-the-art deep neural networks (DNNs) have hundreds of millions of parameters. Due to multiple fully-connected layers, DNNs are compute-intensive and memory-intensive, making them hard to deploy on embedded devices with limited power budgets and hardware resources. Therefore, this paper presents a deep belief network accelerator based on multi-FPGA. Two different schemes, the division between layers (DBL) and the division inside layers (DIL), are adopted to map the DBN to the multi-FPGA system. Experimental results demonstrate that the accelerator can achieve 4.24x (DBL) -6.20x (DIL) speedup comparing to the Intel Core i7 CPU and save 119x (DBL) -90x (DIL) power consumption comparing to the Tesla K40C GPU.

Mermaid: Integrating Vertex-Centric with Edge-Centric for Real-World Graph Processing

There has been increasing interests in processing large-scale real-world graphs, and recently many graph systems have been proposed. Vertex-centric GAS (Gather-Apply-Scatter) and Edge-centric GAS are two graph computation models being widely adopted, and existing graph analytics systems commonly follow only one computation model, which is not the best choice for real-world graph processing. In fact, vertex degrees in real-world graphs often obey skewed power-law distributions: most vertices have relatively few neighbors while a few have many neighbors. We observe that vertex-centric GAS for high-degree vertices and edge-centric GAS for low-degree vertices is a much better choice for real-world graph processing. In this paper, we present Mermaid, a system for processing large-scale real-world graphs on a single machine. Mermaid skillfully integrates vertex-centric GAS with edge-centric GAS through a novel vertex-mapping mechanism, and supports streamlined graph processing. On a total of 6 practical natural graph processing tasks, we demonstrate that, on average, Mermaid achieves 1.83x better performance than the state-of-the-art graph system on a single machine.