Most projections for high-performance, massivelyparallel processors (MPPs) include deep and complex memory hierarchies. Making efficient use of these systems will require making efficient use of these memory hierarch...
详细信息
Recent advances in Internet connectivity and implementations of safer distributed computing through languages such as Java provide the foundation for transforming computing resources into tradable commodities. We have...
详细信息
ISBN:
(纸本)0818684275
Recent advances in Internet connectivity and implementations of safer distributed computing through languages such as Java provide the foundation for transforming computing resources into tradable commodities. We have developed Javelin, a Java-based prototype of a globally distributed heterogeneous, high-performance computational infrastructure that conveniently enables rapid execution of massivelyparallel applications. Our infrastructure consists of three entities: Hosts, clients, and brokers. Our goal is to allow users to buy and sell computational power using supply and demand, and market mechanisms to marshal computational power far beyond what can be achieved via conventional techniques. Several research issues must be worked out to make this vision a reality: allocating resources between computational objects via market mechanisms;expressing and enforcing scheduling and quality of service constraints;modeling programming in a global computing ecosystem;supporting heterogeneous execution without sacrificing computational speed;ensuring host security;global naming and communication;and client privacy.
parallelprogramming needs a high level programming model in which compilers and run time supports take care of traditionally intractable problems related to efficient usage of the target machine (mapping, scheduling,...
详细信息
ISBN:
(纸本)0818684275
parallelprogramming needs a high level programming model in which compilers and run time supports take care of traditionally intractable problems related to efficient usage of the target machine (mapping, scheduling, data decomposition, etc.). The matter of designing a real system providing such a model is highly simplified by constructing the parallel programs using scalable skeletons which capture common structural components of parallel computations. The key problem is the efficient implementation of programs composed of several nested skeleton instances. This requires optimizing the resulting process graph structure and map it on the available resources in order to balance load and minimize communications. The paper describes how this can be done, despite of the intractability of the problems involved, exploiting the 'structure' imposed by the skeleton approach.
Recently, commercial parallel applications became important in parallel computing with the increase of parallel computer users. In this study, we parallelize two commercial applications, a fulltext search system and a...
详细信息
ISBN:
(纸本)0818684275
Recently, commercial parallel applications became important in parallel computing with the increase of parallel computer users. In this study, we parallelize two commercial applications, a fulltext search system and a data mining system. In this paper;the implementation of each application and its performance evaluation on the AP3000 parallel computer are shown. A parallel framework, a novel parallelizing approach applied to those applications, is also described.
We discuss a design principle of massivelyparallel distributed-memory multiprocessor architecture which solves latency problem, and present the Datarol machine architecture. Latencies, caused by remote memory access ...
详细信息
ISBN:
(纸本)0818684275
We discuss a design principle of massivelyparallel distributed-memory multiprocessor architecture which solves latency problem, and present the Datarol machine architecture. Latencies, caused by remote memory access and remote procedure call, are most serious problems in massivelyparallel computers. In order to eliminate the processor idle times caused by these latencies, processors must perform fast context switching among fine-grain concurrent processes. First, we present a processor architecture, called Datarol-II, that promotes efficient fine-grain multi-thread execution by performing fast context switching among fine-grain concurrent processes. In the Datarol-II processor, an implicit register load/store mechanism is embedded in the execution pipeline in order to reduce memory access overhead caused by context switching. In order to reduce local memory access latency, a two-level hierarchical memory system and a load control mechanism are also introduced. Then, me present a cost-effective design of the Datarol-II processor, which incorporates off-the-shelf high-end microprocessor while preserving the fine-grain dataflow concept. The off-the-shelf microprocessor Pentium is used for its core processing, and a co-processor called FMP (Fine-grain Message Processor) is designed for fine grained message handling and communication controls. The co-processor FMP is designed on the basis of FMD (Fine-grain Message Driven) execution model, in which fine-grain multi-threaded execution is driven and controlled by simple fine-grain message communications.
The compilation of high-level programming languages for parallel machines faces two challenges: maximizing data/process locality and balancing load. No solutions for the general case are known that solve both problems...
详细信息
parallel computers with distributed memory are gaining popularity on account of their optimal scalability. However, their efficient use requires a locality-preserving mapping of the application's underlying graph ...
详细信息
parallel computers with distributed memory are gaining popularity on account of their optimal scalability. However, their efficient use requires a locality-preserving mapping of the application's underlying graph structure onto the physical topology of the target platform. PROMOTER is a parallelprogramming model which supports an automatic mapping by the compiler by making the graph structures explicit and thus processable by the implementation. This article describes how this is done for applications with irregular and dynamic spatial structures.
The basic processing element, from PCs to large systems, is rapidly becoming a symmetric multiprocessor (SMP). As a result, the nodes of a parallel computer will often be an SMP. The resulting mixed hardware models (c...
详细信息
The basic processing element, from PCs to large systems, is rapidly becoming a symmetric multiprocessor (SMP). As a result, the nodes of a parallel computer will often be an SMP. The resulting mixed hardware models (combining shared-memory and distributed memory) provide a challenge to system software developers to provide users with programmingmodels that are portable, understandable, and efficient. This paper describes and compares a variety of programmingmodels for a parallel computer made up of SMP nodes.
massivelyparallel Systems (MPSs) with their scalable computation and storage space promises are becoming increasingly important for high-performance computing. The growing acceptance of MPSs in academia is clearly ap...
详细信息
ISBN:
(数字)9783034885348
ISBN:
(纸本)9783764350901;9783034896689
massivelyparallel Systems (MPSs) with their scalable computation and storage space promises are becoming increasingly important for high-performance computing. The growing acceptance of MPSs in academia is clearly apparent. However, in industrial companies, their usage remains low. The programming of MPSs is still the big obstacle, and solving this software problem is sometimes referred to as one of the most challenging tasks of the 1990's. The 1994 workingconference on "programming Environments for massivelyparallel Systems" was the latest event of the working group WG 10.3 of the International Federation for Information Processing (IFIP) in this field. It succeeded the 1992 conference in Edinburgh on "programming Environments for parallel Computing". The research and development work discussed at the conference addresses the entire spectrum of software problems including virtual machines which are less cumbersome to program; more convenient programmingmodels; advanced programming languages, and especially more sophisticated programming tools; but also algorithms and applications.
For serial computation, the object-oriented methodology (O-O) has been shown to aid program modeling, increase reusability, and result in more robust programs. Because of its object-centricity, O-O seems well suited t...
详细信息
ISBN:
(纸本)0818626720
For serial computation, the object-oriented methodology (O-O) has been shown to aid program modeling, increase reusability, and result in more robust programs. Because of its object-centricity, O-O seems well suited to data parallelmodels of massivelyparallelprogramming. Many of the benefits of O-O stem from the arbitrary combining of objects, and the resulting arbitrary message passing patterns. Unfortunately, when working with tens of thousands of processors in parallel, this arbitrariness can result in communication conflicts. This paper proposes a model of objects and communication that resolves this problem.
暂无评论