This paper presents a helper thread prefetching scheme that is designed to work on loosely-coupled processors, such as in a standard chip multiprocessor (CMP) system or an intelligent memory system. Loosely-coupled pr...
详细信息
ISBN:
(纸本)9781424400546
This paper presents a helper thread prefetching scheme that is designed to work on loosely-coupled processors, such as in a standard chip multiprocessor (CMP) system or an intelligent memory system. Loosely-coupled processors have an advantage in that fine-grain resources, such as processor and L1 cache resources, are not contended by the application and helper threads, hence preserving the speed of the application. However, inter-processor communication is expensive in such a system. We present techniques to alleviate this. Our approach exploits large loop-based code regions and is based on a new synchronization mechanism between the application and helper threads. This mechanism precisely controls how far ahead the execution of the helper thread can be with respect to the application thread. We found that this is important in ensuring prefetching timeliness and avoiding cache pollution. To demonstrate that prefetching in a loosely-coupled system can be done effectively, we evaluate our prefetching in a standard, unmodified CMP system, and in an intelligent memory system where a simple processor in memory executes the helper thread. Evaluating our scheme with nine memory-intensive applications with the memory processor in DRAM achieves an average speedup of 1.25. Moreover, our scheme works well in combination with a conventional processor-side sequential L1 prefetcher, resulting in an average speedup of 1.31. In a standard CMP, the scheme achieves an average speedup of 1.33.
Although design pattern is quite useful in software reuse, there are still many barriers when instantiating the design patterns, such as pattern overlapping, traceability, and difficulties in reusing the pattern code....
详细信息
Although design pattern is quite useful in software reuse, there are still many barriers when instantiating the design patterns, such as pattern overlapping, traceability, and difficulties in reusing the pattern code. A role-based approach for design pattern modeling and implementation is proposed. In this approach, roles of pattern are treated as the independent modeling elements and the RoleOf relationship is used to associate a role with an application class. This can improve the reusability of pattern. The meta-model of the RoleOf relationship for pattern instantiation and its semantics are proposed using UML extension mechanism. The stereotypes and tagged values used for identifying pattern information are provided, and it resolves the traceability and overlapping problem in pattern instantiation. The dynamic binding of application and role logic are implemented through the extension to Java language, called Rava. The approach proposed in this paper can effectively solve the problem such as pattern overlapping and traceability during the pattern instantiation, which improves the reusability of pattern logic and guides the software development using design patterns.
At the core of MDD (model driven development) are the concepts of model and its transformation and refinement. Unified modeling language (UML) is selected by object management group (OMG) as a standard modeling langua...
详细信息
At the core of MDD (model driven development) are the concepts of model and its transformation and refinement. Unified modeling language (UML) is selected by object management group (OMG) as a standard modeling language and model driven architecture (MDA) is constructed on it. However, UML models are not precisely described, especially in semantics. Thus the models developed in different phases or constructed in different views are not easily to be integrated together in MDA. In this paper, based on Hoare and He's unifying theories of programming (UTP), a method is proposed to combine the refinement calculus of object systems (rCOS) with UML is proposed to increase the precision and transformation ability of the models. Models at different abstract levels and different views are constructed and integrated together to form a unified modeling system.
The first-come first-served scheduling framework adopted by most application servers has been proved to be inappropriate for dealing with unexpected overload for Internet-oriented Web applications. Considering the def...
详细信息
The first-come first-served scheduling framework adopted by most application servers has been proved to be inappropriate for dealing with unexpected overload for Internet-oriented Web applications. Considering the deficiency of the existing work from the architecture view, this paper presents a scheduling framework based on the notion of QoS benefits, which contains several cooperating components to guarantee the QoS requirements of the applications. The QoS benefits used to evaluate the QoS guarantee provided by the server according to the QoS requirements of the applications and the resource management based on the QoS benefits will help to provide a better QoS guarantee. Experimental results confirm the efforts on the OnceAS application Server.
This paper presents a failure diagnosing method based on the combinatorial design for testing. This method analyzes the schemas included by the correct running test cases and failure caused test cases in the combinato...
详细信息
This paper presents a failure diagnosing method based on the combinatorial design for testing. This method analyzes the schemas included by the correct running test cases and failure caused test cases in the combinatorial testing suite, it then can conclude that the errors must be in a very small range through analyzing the test cases and retesting with some complementary test cases. So it can provide the very efficient and valuable guidance for the debugging and testing of software. This paper gives a further study on Combinatorial Design Approach for Testing, which is studied and applied widely for its scientificity and effectiveness in software testing with a quite small test suite, especially for the software under testing whose faults come mainly from the parameters or the interactions of the system parameters.
Automated simplification of trigonometric expressions is an important problem that hasn't been completely solved by current computer algebraic systems. Even if the symbolic computation softwares like Maple, Mathem...
详细信息
Automated simplification of trigonometric expressions is an important problem that hasn't been completely solved by current computer algebraic systems. Even if the symbolic computation softwares like Maple, Mathematica and Derive, etc., they are still not applicable enough to the problem. This paper presents a number of unique prescriptions for the ordering of some trigonometric rules, which have been derived by observing how human experts follow their intuitive rules. The authors have implemented the procedure in Lisp because of its suitability for formula manipulations and rule-based reasoning systems. Consequently, it can simplify many trigonometric expressions that are even difficult to do by hand, simultaneously with readable process, and it achieves much better results for many hard problems than Maple and Mathematica do.
An authenticated group key agreement protocol suite (PAGKA) based on pairings is presented in this paper. This suite extends an efficient Diffie-Hellman-based protocol suite by introducing parings on elliptic curves a...
详细信息
An authenticated group key agreement protocol suite (PAGKA) based on pairings is presented in this paper. This suite extends an efficient Diffie-Hellman-based protocol suite by introducing parings on elliptic curves and public-key certificates. The result suite builds its security on the hardness of the bilinear Diffie-Hellman (BDH) problem in the random oracle model, and provides implicit key authentication, perfect forward secrecy and unknown key-share secrecy.
Network partitioning method is widely used in VLSI design or load balancing. Iterative partitioning algorithm such as the Fiduccia and Mattheyses' algorithm or the Sanchis' algorithm uses local information and...
详细信息
ISBN:
(纸本)9781932415582
Network partitioning method is widely used in VLSI design or load balancing. Iterative partitioning algorithm such as the Fiduccia and Mattheyses' algorithm or the Sanchis' algorithm uses local information and have tendency to fall in local optima. CLIP is an iterative algorithm that uses cluster information in order to overcome local optima cases. We uses the concept of CLIP and enhance the cluster information to find clusters more exactly. Main idea of our proposed algorithm is to compel the cell movement direction until a cluster is identified. The result of the proposed algorithm shows better results against CLIP when a network has a big cell.
ASIC design in the area of image processing needs large input output pixel data to be dealt with. The input image data is generated from input image which creates problems during ASIC simulation process. Manual manipu...
详细信息
暂无评论