The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
The solutions to today's foremost scientific challenges require order-of-magnitude increases in computing power. The route to TeraFLOP computing lies in parallel multi-computers that exploit advances in microprocessor technology. As the only manufacturer of both advanced microprocessors and parallel supercomputers, Intel Corp. plays a unique role in the drive toward TeraFLOP computing. The...
The Inmos transputer family is a range of VLSI components for concurrent processing systems. Transputers are used in applications such as communications, real-time control, image processing, robotics, databases, graphics and animation. New, software compatible, transputer products will exploit parallelism in instruction execution to provide greatly increased processor performance. New communications...
A virtual shared memory architecture (VSMA) is a distributed memory architecture that looks to the application software as if it were a shared memory system. The major problem with such a system is to maintain the coherence of the distributed data entities. Shared virtual memory means that the shared data entities are pages of local virtual memories with demand paging. Memory coherence may be strong...
The parallel language EDS Lisp has been designed for distributed memory machines. In this respect, it presupposes only a minimal required coherency in the storage view of parallel processes. This paper describes the coherency model of EDS Lisp and shows how the requirements of EDS Lisp can be realized using a weak coherency model on the parallel distributed memory EDS machine.
In this paper we have described a user-definable memory coherence scheme for distributed shared memory that is flexible enough to meet the varying needs of a wide variety of user applications. We believe that the concepts presented in this paper will be useful for the design of other distributed systems.
This paper describes the parallelization of Multigrid Programs with SUPERB. SUPERB is an interactive SIMD/MIMD parallelizing system for the SUPRENUM machine. We present multigrid oriented features of our data partitioning language and discuss interprocedural aspects in the parallelization of these programs. One of our results is that interactivity is inherent to the parallelization of these programs.
As we search for a mapping that will avoid the routing of data, our interest in this paper was to characterize the graphs for which there exits a topological mapping onto an RTRN and to simulate the behaviour of an application running onto this machine according to this mapping. After we have determined the kind of graph which is assignable we realize a software tool which simulates the execution...
This paper describes SIMPLE: a performance evaluation tool environment for parallel and distributed systems based on monitoring of concurrent interdependent activities. We emphasize the tool environment as a prerequisite for successful performance evaluation. All tools use the data access interface TDL/POET which can decode measured data of arbitrary structure, format and representation. This makes...
In this paper we present a systematic method for mapping systolizable problems onto Distributed Memory Multiprocessors. A systolizable problem is a problem for which it is possible to design a Systolic Algorithm. As a first stage, we design a Systolic Algorithm for the problem to be solved. Then, different transformations are applied to this Systolic Algorithm in order to adapt it to the available...
In this paper, a partial ordering of synchronization events for the debugging of distributed programs in tightly-coupled multiprocessor systems is defined. Techniques for the debugging of parallel programs require timestamping of events. The physical clocks of different processors are in general not suitable for time stamping events, because these clocks are not synchronized. Synchronization of the...
Most parallel applications in a Transputer-network require a lot of communication between the processing nodes. For such applications the communication system TRACOS was developed to support data transfer between arbitrary Transputers in the network. To maximize the performance of the parallel system its dynamic internal behavior has to be analyzed. For this purpose event-driven monitoring is an appropriate...
A breakthrough of distributed memory architectures as general purpose programmable computers depends strongly on the ease of programming such machines. TOPSYS (TOols for Parallel SYStems) is an integrated environment for programming distributed memory multiprocessors. Its concepts and application on an iPSC/2 multiprocessor are presented. It is based on a simple message passing process model implemented...
We present an optimal all-to-all communication procedure for a mesh connected topology of processors with wrap-around (a torus). This procedure is based on a spanning tree method and uses a reduced bufferization space.
We describe a tool that supports data-parallel programming on a distributed memory multiprocessor. Load balancing, based on an evolution algorithm, is integrated into this environment. Timing results for balanced Jacobi relaxation on an adaptive unstructured grid show that the encurred overhead is acceptable for calculation intensive applications.
This paper presents an object-oriented language interface as a basis for the programming of loosely-coupled multiprocessor systems and an algorithm for automatic translation into parallel programs. Starting with a sequential program written in a restricted subset of C++, the translator algorithm first generates a machine-independent communication graph and proceeds with the creation of the machine-specific...
In this paper we present a unified approach for compiling programs for Distributed-Memory Multiprocessors (DMM). Parallelization of sequential programs for DMM is much more difficult to achieve than for shared memory systems due to the exclusive local memory of each Virtual Processor (VP). The approach presented distributes computations among VPs of the system and maps data onto their private memories...
Distributed-Memory, point-to-point connected MIMD-architectures (multicomputers) realize the communication between their constituent nodes by message-passing interconnection networks. The development of high-speed cut-through-type routing modules for interconnection networks has passed two generations: modules of the first generation offer distributed applications a pseudo-complete interconnection...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.