Chia-Ying Tseng

The Design of Way-Prediction Scheme in Set-Associative Cache for Energy Efficient Embedded System

Embedded System develops rapidly, functions turn into more complicate, and multi-media applications are growing daily and they consume more electrical power. Therefore, how to improve stand-by time will become a very important issue. Related researches indicate that the power consumption of processor cache is accounted for a big proportion. Way-prediction and LRU (Least Recently Used) algorithms improve hit rate and would help in reducing the number of tag comparisons, and therefore save energy consumption. In this paper, we use MRU (Most Recently Used) table to record the most used block for each index and use Modified Pseudo LRU (MPLRU) Replacement algorithm for reducing hardware complexity and cache miss rate.Experiments show our prediction hit rate reach 90.15%, thus save 64.12% energy. The experimental results are obtained by using Wattch cache simulator for SPEC95 benchmarks.

An extenics-based load balancing mechanism for distributed computing systems

In distributed computing systems, load balancing is one of the most important factors that affect the system performance. This paper presents sender-initiated dynamic load balancing policies in a distributed computing system. There are four policies consisted in the proposed extenics-based load. balancing mechanism (ELBM). For the transfer policy, instead of defining a fixed threshold value in the traditional load balancing policies, we apply the extension set theory and use the average response time of jobs as a factor to create an adaptive threshold value which is calculated by a relational function of average response time, for determining that a new arriving job should be migrated or not. In addition to the transfer policy, a bypass-transfer method is used in the location policy. Jobs that need to be migrated do not need to be transferred to central node first, but are transferred to destination node directly. so as to reduce the communication costs. In order to make the system more reliable, a template queue is added at the master node of the system to reduce the job arrival failure. A simulation model has been built for evaluating the performance of the system. Compared with other load balancing mechanisms, such as RT and ALBCII algorithms, the simulation results show that a better performance can be achieved by the proposed mechanism.

Dynamic Dispatching Tasks Management for H.264 Encoder on Heterogeneous Dual-Core Embedded System

This paper shows the mechanism of using a dual-core embedded system to process the multimedia data and how the resources managed to serve the whole process. The partitioning methodologies now for asymmetric dual-core systems typically perform software partitioning during compile time, which assigns tasks simply to RISC or DSP only, leaving the other processor idle. If we can use both processors to handle part of the job, the whole process will be faster and save a lot of time and power. In this paper we proposed a dynamic dispatching method to manage and schedule the tasks between RISC and DSP. We dispatch the tasks to both RISC and DSP core. We use TI’s DaVinci DM6446 as our example to show the enhancement of system performance. The paper also indicates the experiment results and the further issues to promote.

A Dynamic Reservation and Call Admission Control Policy with HMM for Multimedia Cellular Networks

The multimedia services are getting to become the major trend in next-generation cellular networks. Call admission control (CAC) plays the key role for guaranteeing the quality of service (QoS) in cellular networks. The goal which keeps both the call dropping probability (CDP) and call blocking probability (CBP) as low as possible is more difficult owing to the user indeterminate mobility. In this paper, the hidden Markov models (HMM) concept which is suitable for solving a dynamic situation is introduced and applied to the call admission control policy. The prediction of user mobility can be modeled and resolved as a decoding problem for the HMM. According to the prediction result, the proposed CAC method can reserve appropriate bandwidths for a handoff call beforehand, thus the call dropping probability is almost the same as the traditional guard channel policies that can keep lower CDP than unpredicted ones, moreover, the call blocking probability is also lower than the traditional guard channel policies since the proposed method can reserve the suitable bandwidths in the appropriate cells but not reserve a stationary ones which are always adopted by traditional CAC methods. Therefore, the proposed method can not only satisfy both CDP and CBP issues, but also improve the system utilization

On implementation of a scalable wallet-size cluster computing system for multimedia applications

In order to obtain a low-cost, low-power consumption and small sized digital home processing center, especially for multimedia applications, we propose a scalable wallet-size cluster computing system in this paper. Such system uses low-cost, low-power consumption and small sized microprocessors as computing nodes to form a cluster system. The main reason why we are interested to investigate and build such system is to increase the computing power and to accelerate the H.264 video processing. We analyzed the H.264 decoder scheme and distributed the most heavy-load computations in the decoder scheme to each computing node (or microprocessor) of the cluster system for the processing of signals, in parallel. To run multimedia applications in this cluster system, an efficient real-time video processing is implemented.

An adaptive scheduler for embedded multi-processor real-time systems

In the real-time system, the system usually supports the application that includes the task set with hard and soft real-time tasks. The traditional researches on the task scheduling in real-time operating systems were setting the priority for providing service to soft tasks after that can promise meet deadline of hard tasks. Therefore, those scheduling schemes were trying to provide better average meet ratio for soft tasks, but did not consider the deadline of each task. In this paper, we propose an adaptive scheduler based on the global scheduling. This scheduler can schedule a set of periodic tasks which composed of hard and soft and increase the average meet ratio for soft tasks. The proposed adaptive scheduler for the real-time multiprocessor systems considers the operating environment in the real world. The simulation results show that the proposed adaptive scheduler can improve the system performance and fit for almost all real-time applications.

The adaptive layer-based scheduling system for embedded real-time transmission on scalable multimedia stream

With the recent development in compression and network technology, stream media has been adopted in internet and intranet. Application of stream media technology will play a key role in future development of fast networks. Original video/audio data will be stored in the devices for storage after having been pre-compressed by a video/audio compression algorithm. When clients have requirements, the stream server retrieves the video/audio data from the storage devices through the network. This paper discusses a simple and effective real-time scheduling algorithm, adaptive Layer-Based Least-Laxity-First (LB-LLF) scheduling algorithm, to improve the output quality of video/audio on a network and to achieve synchronised playback effect. The proposed algorithm considered real-time constraint, unequal priorities of scalable media stream in different layers, and a good tradeoff between coding efficiency and drifting error. This guarantees the effective usage of available channel bandwidth and the better quality of playback in client.

A Wallet-Size Cluster for H.264 Encoding

A scalable wallet-size cluster computing system has been proposed to obtain low cost, low power consumption and small sized digital home processing center. In this paper, we apply the wallet-size cluster for H.264 encoding to accelerate the H.264 video processing. We analyze the encoder scheme of the H.264, and perform the H.264 encoding in parallel through the nodes of the cluster. The experimental results show that an efficient real time video processing for H.264 can be achieved