A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....Parallel computing on clusters of workstations and personal computers has very high potential, since it leverages existing hardware and software. In fact, there are a number of existing commercial parallel programming environments that can run on top of clusters of workstations [3, 11, 15, 18] Parallel programming environments offer the user a convenient way to express parallel computation and communication. The communication part consists of the usual point to point communication as well as collective communication. Examples ot collective communication operations include one to all ....

A. Skjellum and A. P. Leung, "Zipcode: a portable multicomputer communication library atop the reactive kernel", Proceedings of the 5th Distributed Memory Computing Conference, IEEE, pp. 328-337 April 1990.

....than selecting one of them and adopting it as the standard. Thus, MPI has been strongly influenced by work at the IBM T. J. Watson Research Center [1, 2] Intel s NX 2 [23] Express [22] nCUBE s Vertex [21] p4 [7, 6] and PARMACS [5, 8] Other important contributions have come from Zipcode [24, 25], Chimp [14, 15] PVM [4, 11] Chameleon [19] and PICL [18] The M Pl standardization effort involved about 60 people from 40 organizations mainly from the United States and Europe. Most of the major vendors of concurrent computers were involved in M P I, along with researchers from ....

....5.1.2 MPl s Support for Libraries Tile corresponding concepts that M Pl provides, specifically to support robust libraries, are as follows: Contexts of communication, Groups of processes, Virtual topologies, Attribute caching, Communicators. Communicators (see [16, 24, 27]) encapsulate all of these ideas in order to provide the appropriate scope for all communication operations in M PI. Communicators are divided into two kinds: intra communicators for operations within a single group of processes, and inter communicators, for point to point communication between ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767 776. IEEE Press, 1990.

....than selecting one of them and adopting it as the standard. Thus, MPI has been strongly influenced by work at the IBM T. J. Watson Research Center [1, 2] Intel s NX 2 [23] Express [22] nCUBE s Vertex [21] p4 [7, 6] and PARMACS [5, 8] Other important contributions have come from Zipcode [24, 25], Chimp [14, 15] PVM [4, 11] Chameleon [19] and PICL [18] The MPI standardization effort involved about 60 people from 40 organizations mainly from the United States and Europe. Most of the major vendors of concurrent computers were involved in MPI, along with researchers from universities, ....

....133 5.1.2 MPI s Support for Libraries The corresponding concepts that MPI provides, specifically to support robust libraries, are as follows: ffl Contexts of communication, ffl Groups of processes, ffl Virtual topologies, ffl Attribute caching, ffl Communicators. Communicators (see [16, 24, 27]) encapsulate all of these ideas in order to provide the appropriate scope for all communication operations in MPI. Communicators are divided into two kinds: intra communicators for operations within a single group of processes, and inter communicators, for point to point communication between two ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....including all those present were held. Many vendors of concurrent computers were involved, as were researchers from universities, government laboratories, and industry. This effort culminated in the 1994 publication of the MPI specification [8] Other sources of information on MPI are available [10] or are under development (see the sidebar, More MPI Assistance ) Researchers incorporated into MPI the Jack J. Dongarra, Steve W. Otto, Marc Snir, and David Walker Standard MPP and Work st at ions for COMMUNICATIONS OF THE ACM July 1996 Vol. 39, No. 7 85 most useful features of several ....

....Steve W. Otto, Marc Snir, and David Walker Standard MPP and Work st at ions for COMMUNICATIONS OF THE ACM July 1996 Vol. 39, No. 7 85 most useful features of several systems, rather than choosing a single system as the standard. MPI has roots in PVM [4, 5] Express [9] P4 [1] Zipcode [10], and PARMACS [2] and in systems sold by IBM, Intel, Meiko Scientific, Cray Research, and nCube. MPI is used to specify the communication among a set of processes forming a concurrent program. The message passing paradigm is attractive because of its wide portability and scalability. It is ....

Skjellum, A., and Leung, A. Zipcode: A portable multicomputer communication library atop the reactive kernel. In Proceedings of the 5th Distributed Memory Concurrent Computing Conference, D.W. Walker and Q.F. Stout, eds. (Charleston, S. Car, Apr. 1990) IEEE Press, 1990, pp. 767--776.

....Intel s NX [Pierce 88] and NCUBE s Vertex [NCUBE 87] are examples. Other packages are implemented as user level libraries; these include Thinking Machines CMMD [TMC 92] Parasoft s Express [Parasoft 88] and the public domain packages PVM [Sunderam 92] PICL [Geist et al. 91] and Zipcode [Skjellum Leung 90] These packages have slightly different purposes and features, but in most ways they are very similar. In response to the proliferation of message passing packages, a standardization effort was launched. The proposed standard, called MPI1, is meant to capture the established features common ....

A. Skjellum and A. Leung. Zipcode: A portable multicomputer communication library atop the reactive kernel. In Proceedings of Fifth Distributed Memory Computing Conference, pages 767--776. IEEE Press, 1990.

....into a programming notation is not discussed, and the notion of virtual computer is absent. Some recent work on parallel message passing libraries has explored the use of process groups and communication contexts to support the encapsulation of communication operations in parallel libraries [28, 2]. Chien and Dally s Concurrent Aggregates (CA) language allows the definition of homogeneous collections of objects called aggregates; messages addressed to an aggregate are routed to one of its members [12] As in this paper, concurrent structures can be defined and composed with other structures ....

A. Skjellum and A. Leung, "Zipcode: A portable multicomputer communication library atop the Reactive Kernel," in Proc. 5th Distributed Memory Concurrent Computing Conf., IEEE Press, 1990, pp. 767--776.

....widely used and well understood paradigm for the implementation of parallel programs on distributed memory architectures is that of message passing. Several message passing libraries are available in the public domain, including p4 [8] Parallel Virtual Machine (PVM) 9] PICL [10] and Zipcode [11]. Recently, a core of library routines (influenced strongly by existing libraries) has been standardized in the Message Passing Interface (MPI) 12,13] Public domain implementations of MPI are widely available for parallel hardware and clusters of workstations, and most high performance computing ....

A. Skjellum and A. Leung, "Zipcode: A portable multicomputer communication library atop the reactive kernel," in Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pp. 767--776, IEEE Press, 1990.

....1.1.2 MPI s Support for Libraries The corresponding concepts that MPI provides, specifically to support robust libraries, are as follows: ffl Contexts of communication, ffl Groups of processes, ffl Virtual topologies, ffl Attribute caching, ffl Communicators. Communicators (see [2, 3, 5]) encapsulate all of these ideas in order to provide the appropriate scope for all communication operations in MPI. Communicators are divided into two kinds: intra communicators for operations within a single group of processes, and intercommunicators, for point to point communication between two ....

.... IN ranges an array of integer triplets, of the form (first rank, last rank, stride) indicating ranks in group of processes to be included in newgroup OUT newgroup new group derived from above, in the order defined by ranges (handle) int MPI Group range incl(MPI Group group, int n, int ranges[ [3], MPI Group newgroup) MPI GROUP RANGE INCL(GROUP, N, RANGES, NEWGROUP, IERROR) INTEGER GROUP, N, RANGES(3, NEWGROUP, IERROR If ranges consist of the triplets (first 1 ; last 1 ; stride 1 ) first n ; last n ; stride n ) then newgroup consists of the sequence of processes in group ....

[Article contains additional citation context not shown here]

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....rather than selecting one of them and adopting it as the standard. Thus, MPI has been strongly influenced by work at the IBM T. J. Watson Research Center [1, 2] Intel s NX 2 [16] Express [15] nCUBE s Vertex [13] P4 [3] and PARMACS [11, 12] Other important contributions have come from Zipcode [17, 18], Chimp [7, 8] PVM [9, 19] and PICL [10] One of the objectives of this paper is to promote a discussion within the concurrent computing research community of the issues that must be addressed in establishing a practical, portable, and flexible standard for message passing. This cooperative ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....the library to implement the necessary data movement among processors for HPF REDISTRIBUTE. The primary advantages to using MPI are its portability, ease of use, and in particular, its support for parallel library development [12] MPI uses communication contexts, first proposed in Zipcode [13], to provide safe communication spaces for processes. Communication contexts shield processes from unrelated ongoing communication which otherwise may be interpreted, erroneously, as related to the current communication due to the asynchronous behavior of SPMD programs. This is particularly ....

A. Skjellum and A. P. Leung, "Zipcode: A portable multicomputer communication library atop the reactive kernel," in Proceedings of the Fifth Distributed Memory Computing Conference (DMCC5), pp. 767--776, Apr. 1990.

....them and adopting it as the standard. Thus, MPI1 has been strongly influenced by work at the IBM T. J. Watson Research Center by Bala, Kipnis, Snir and colleagues [1,2] Intel s NX 2 [17] Express [16] nCUBE s Vertex [14] and PARMACS [11,12] Other important contributions have come from Zipcode [18,19], Chimp [6,7] PVM [8,20] and PICL [9] One of the objectives of this paper is to promote a discussion within the concurrent computing research community of the issues that must be addressed in establishing a practical, portable, and flexible standard for message passing. This cooperative process ....

....phase. This situation can arise if the message selectivity criteria of the two phases overlap, as may be the case when using a canned concurrent software library in which the selectivity criteria, in general, are unknown. Communication contexts, first used in the Zipcode message passing system [18,19], provide a means of disambiguating such situations. In effect, a communication context provides a third selectivity criterion, in addition to type and source process, that may be used to control the receipt of messages. A communication context is uniquely labeled by a strictly positive integer ....

A. Skjellum and A. Leung, "Zipcode: a Portable Multicomputer Communication Library atop the Reactive Kernel," Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pp. 767--776 (IEEE Press 1990).

....communication routines that can be invoked from within FORTRAN or C application programs. Over the past several years, a large number of programming environments and communication libraries for parallel computers have been developed, including PVM [3] LINDA [5] PICL [9] PARMACS [11] Zipcode [13], Express [15] the NCUBE 2 library [12] the CM 5 library [14] and the iPSC 860 library [16] The design of the EUI borrows some of the popular communication concepts that already exist in many of these libraries, and in addition, provides several novel operations. The primary goal in the design ....

A. Skjellum and A. P. Leung, "Zipcode: a portable multicomputer communication library atop the Reactive Kernel", Proceedings of the 5th Distributed Memory Computing Conference, IEEE, April 1990.

.... by widely used message passing systems such as Express [12] NX 2 [13] Vertex, 11] parmacs [8, 9] and P4 [10] In addition, the flexibility and usefulness of mpi has been broadened by incorporating ideas from more recent and innovative message passing systems such as chimp [4, 5] Zipcode [14, 15], and the IBM External User Interface [7] The general design philosophy followed by mpi is that while it would be imprudent to include new and untested features in the standard, concepts that have been tested in a research environment should be considered for inclusion. Many of the features in ....

....routine call. In this case the program behaves as intended. Note that the second message sent by process 2 is received by process 0, and that the message sent by process 0 is received by process 2. 3.1. 2 Communication Contexts Communication contexts, first used in the Zipcode communication system [14, 15], promote software modularity by allowing the construction of independent communication streams between processes, thereby ensuring that messages sent in one phase of an application are not incorrectly intercepted by another phase. Communication contexts are particularly important in allowing ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....value is not legal on a particular machine. Programs written using PICL routines will also produce timestamped trace data on interprocessor communication, processor busy idle times and simple user defined events if a few additional statements are added to the source code. ffl ZIPCODE: ZIPCODE [99] is a portable message passing system which includes a number of higher level features which are necessary for building libraries and large scale application software. These include process groups, communication contexts, collective operations, and virtual topologies. 4.1.5 P4 P4 (Portable ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communications library atop the Reactive Kernel. In Proceedings of the 5th Distributed Memory Computer Conference, pages 767--776. IEEE Press, 1990.

....1 Introduction Parallel computing on clusters of workstations and personal computers has very high potential, since it leverages existing hardware and software. In fact, there are a number of existing commercial parallel programming environments that can run on top of clusters of workstations [4, 11, 15, 18, 20, 21, 22]. Parallel programming environments offer the user a convenient way to express parallel computation and communication. The communication part consists of the usual point topoint communication as well as collective communication. Examples of collective communication operations include one to all ....

A. Skjellum and A. P. Leung, "Zipcode: a portable multicomputer communication library atop the reactive kernel", Proceedings of the 5th Distributed Memory Computing Conference, IEEE, pp. 328--337, April 1990.

....difficult. First, a direct implementation on MIMD machines is nontrivial. The user must handle the synchronization and load balance issues at the same time, which could be extremely difficult for some application problems. In general, a runtime support system, such as LINDA [2, 5] reactive kernel [27, 33], or chare kernel [30] is necessary for solving asynchronous problems. Implementation of the asynchronous problems on SIMD machines is even more difficult because it needs a runtime support system, and the support system itself is asynchronous. In particular, the support system must arrange the ....

Anthony Skjellum, Alvin P. Leung, and Manfred Morari. Zipcode: A portable multicomputer communication library atop the reactive kernel. In Proc. of the 5th Distributed Memory Computing Conf., pages 767--7765, April 1990.

....interconnection networks and point to point networks with virtual cut through like routing algorithms. In fact, this model has been widely used by several researchers (see [24] and has been adopted by numerous communication libraries, such as Express [21] PARMACS [25] PICL [23] Zipcode [31], Venus [5] and CCL [4] Two complexity measures are frequently used in evaluating communication algorithms in such a model. The first measure, denoted by C 1 , is the number of communication rounds required by an algorithm. Notice that in each round, the messages sent and received by each ....

A. Skjellum and A. P. Leung, Zipcode: a portable multicomputer communication library atop the Reactive Kernel, 5th Distributed Memory Computing Conference, IEEE, April 1990.

....10.00000 17.00000 14.00000 13.00000 10.00000 0.00000 7.00000 21.000000 20.00000 17.00000 7.00000 0. 00000 Figure 4: Example LOAPSP Program Output The PVM system is an evolving research code, No semaphores (could be used with concurrent I O over NFS) No message passing contexts (cf, [15]) No receipt selectivity based on source (type only) Single buffer send receive limitation, Several key issues of distributed computing are not addressed well: 3 Load balancing (none) 3 Fault tolerance (master slave, redundancy, recovery) 3 Scalability (linear spawn ....

Skjellum, A. and A. P. Leung, Zipcode: A Portable Multicomputer Communication Library atop the Reactive Kernel, Proc. of the Fifth Distributed Memory Computing Conf., IEEE Press, pp. 767-776, April 1990.

No context found.

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....are commonly used on many types of machines, ranging from massively parallel supercomputers to clusters of workstations. The reader is assumed here to have a significant knowledge of the de facto MPI 1 standard [14, 15] an effort influenced by over a decade of research and commercial systems [2, 3, 7, 11, 12, 19, 21, 23, 24, 25, 27, 29]. This interface has been realized in several commercial implementations, based on well known public implementations [5, 20] In this paper, we present two object oriented class libraries for supplementing MPI, in support of C . The features and limitations of these systems are indicated, as is ....

A. Skjellum and A. Leung. Zipcode: a portable multicomputer communication library atop the reactive kernel. In D. W. Walker and Q. F. Stout, editors, Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pages 767--776. IEEE Press, 1990.

....divorced from its original relationship with the Reactive Kernel Cosmic Environment (CE RK) 3. 5 Contexts of Communication A communication context is an abstraction that was introduced by the author in the original (1988) Zipcode system, and which also will appear in the MPI1 standard [12, 2]. In order to write practical, safe distributed memory, and or distributed computing libraries, communication contexts are needed to restrict the scope of messages. This is done to prevent messages from being selected improperly by processes when they do message passing. We described contexts in ....

.... safe distributed memory, and or distributed computing libraries, communication contexts are needed to restrict the scope of messages. This is done to prevent messages from being selected improperly by processes when they do message passing. We described contexts in several papers on Zipcode [12, 15, 14]. Without this type of scope restriction, it quickly becomes intractable to build up code without globalizing the details of how each portion of a code utilizes the message passing resource. Communication contexts are therefore central to creating reusable library code, and to maintaining ....

Anthony Skjellum and Alvin P. Leung. Zipcode: A portable multicomputer communication library atop the Reactive Kernel. In Proc. Fifth Distributed Memory Computing Conf. (DMCC5), pages 767--776. IEEE, April 1990.

.... Primarily, the upshot of these requirements includes the necessity of concensus on the mechanisms for each safety (often bit patterns and widths of bit fields) 2 Common Context Support of Interoperable Implementations The classical example of a context supporting messaging system, Zipcode [10, 11], used a global name server for the support of contexts across communicating groups of processes. In MPI, explicit choices were made to eliminate this requirement 3 , and allow local context support, and to make this quantity purely ineffable, so that implementations could achieve communication ....

Anthony Skjellum and Alvin P. Leung. Zipcode: A Portable Multicomputer Communication Library atop the Reactive Kernel. In Proceedings of the Fifth Distributed Memory Computing Conference (DMCC5), pages 767--776. IEEE, April 1990.

....Mississippi State University. 1 Introduction Zipcode was developed in 1988 at the California Institute of Technology by the first author, and was developed further at Lawrence Livermore National Laboratory, with additions and support continuing to this date at Mississippi State University [29] This message passing system was strongly influenced by the point to point semantics of the Reactive Kernel primitives [23, 24, 25] by the collective primitives of CrOS [15, Chapter 14] and by the process management features of the Cosmic Environment [23] all developed at Caltech) Zipcode ....

.... messages) We don t usually reveal the internal representations of process names at the application level (except in some Zipcode primitives) though the Zipcode implementation currently makes specific assumptions about a 64 bit naming convention based on a node and pid pair of integers [29] Thus, groups define a rank naming for processes in point to point communication relative to the group. In addition, groups define the scope of collective operations. This scope allows us to make strong statements about the non interference of sequential collective operations in the same ....

[Article contains additional citation context not shown here]

Anthony Skjellum and Alvin P. Leung. Zipcode: A Portable Multicomputer Communication Library atop the Reactive Kernel. In Proceedings of the Fifth Distributed Memory Computing Conference (DMCC5), pages 767--776. IEEE Press, April 1990.

No context found.

A. Skjellum and A. Leung, "Zipcode: A portable multicomputer communication library atop the reactive kernel," in Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pp. 767--776, IEEE Press, 1990.

No context found.

A. Skjellum and A. Leung, "Zipcode: A portable multicomputer communication library atop the reactive kernel," in Proceedings of the Fifth Distributed Memory Concurrent Computing Conference, pp. 767--776, IEEE Press, 1990.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC