Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HP Java. Concurrency: Practice and Experience, 9(6), June 1997.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....interpreter Java Runtime Just in time Compiler Figure 2.10. Java computing model 21 2.4.1.2. Java Performance We may reach a factor of 2 4 lower performance from a PC JIT compiler compared to C compiled code [30] 2] For the portable applet mode this is as big as a factor of two[12][4]. However with some restrictions on programming style[15] we expect Java language and VM compilers to be competitive with the best Fortran and C compilers [9] 2.4.2. Java Threads Threads are created and managed by the classes Thread and ThreadGroup. Creating a Thread object creates a thread. ....

Carpenter, B., Chang, Y., Fox, G., Leskiw, D., and Li, X. Experiments with hpjava. Concurrency: Practice and Experience 9, 6 (Jun 1997), 633-648.

....and for regular and irregular collective communications. Such classes are useful for JavaParty. The system runs on an IBM SP 2, but in contrast to JavaParty s 100 pure Java approach, it uses a runtime system written in C that interacts with a native MPI communication library. In the HPJava [2] project at Syracuse, Carpenter, Fox et al. use wrappers to interface to native HPF code and library based extensions of Java. They offer useful classes for distribution of array data and for collective communication as well. Whereas JavaParty has the capability of transforming a single program ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with "HPJava". Concurrency: Practice and Experience, June 1997.

....that can be implemented by a class. It consists of one or more function definitions. If a class has to conform to the protocol it must implement all the functions in the interface. 2 Any class that implements the Array interface can be accessed through the normal array access notation (e.g. a[1,2]) Sparse matrices and block matrices are supported by classes that implement this interface. Spar also supports elastic arrays (arrays that can be grown and shrunk during their lifetime) macros, and array statements, but discussion of these features is beyond the scope of this paper. For a more ....

....in the standard Java class library. We feel that for the intended applications (computer intensive, often arraybased computations) this programming model is too inconvenient and too error prone. For similar reasons parallel programming models based on message passing and or SPMD support libraries [1, 2, 11] have not been chosen. Since the data parallel programming model, as used in HPF [6] is restricted to semi regular array computations, this was also not adopted. Instead of all these programming models, we have chosen to provide a more implicit programming model through the language constructs ....

B. Carpenter, Y.-J. Chang, G. Fox, D. Leskiw, and X. Li. Experiments with HP Java. Concurrency: Practice and Experience, 9(6), June 1997.

....the control of the computational process is responsible for a significant amount of the complexity and execution time. In different studies that evaluate the potential of Java as a highperformance computing language, it is indicated that certain improvements are necessary for successful usage [2, 3]. On the other hand, Java offers an efficient and flexible simulation engine for DES with its complex control of the computational process. In general, simulation studies and experiments are pursued with specific simulation languages such as MODSIM III, SIMULA 67 or Pearl [7] or with general ....

.... public void Deliver(int nbytes) f counter; g abstract void Deliver(long t, int nbytes) public void main(Entity peers[ f Server svr = peers[0] long t = new Random( nextInt( for(int i=0; i niter; i ) f svr.Read( bytes[i] 0] svr.Write( bytes[i] 1] svr.Prefetch( bytes[i][2], this ) waitfor( t2 , Deliver) svr.Cancel( g g g Figure 3: The Client entity performs a number of Read, Write and Prefetch requests to the Server entity, to which it is connected by the neighbor list. It counts the number of Prefetch requests which could be satisfied within a period t2 ....

[Article contains additional citation context not shown here]

CARPENTER,B.,CHANG,Y.-J.,FOX,G.,LESKIW, D., AND LI, X. Experiments with `HP Java'. Concurrency: Practice and Experience 9, 6 (June 1997), 633--648.

....C have shown [2] By binding the original native libraries like BLAS, Java programs can gain in performance on all those hardware platforms where the libraries are efficiently implemented. A Java to PVM interface is publicly available [19] A Java binding for a subset of MPI has also been written [4, 7] and run on up to 8 processors of Sun Ultra Sparc. In comparison, our binding has been generated automatically to cover the whole of MPI; it allows for better use of MPI derived data types; its flexibility makes it easier to retarget to different versions of the Java native interface and hardware ....

B. Carpenter, Y-J. Chang, G. Fox, D. Leskiw, and X. Li. Experiments with "HPJava". Concurrency: Practice and Experience, 9(6):633--648, 1997.

....as required for, eg, address translation. This can be seen as a new compiler targetting the run time an HPJava although the work of this compiler is much simpler than for HPF, because the programmer will explicitly specify the communication requirements. For early results of the work on Java see [20, 21, 46, 45, 50]. 9 PCRC Java interface MPI Distributed data and control ad interface (Adlib) Kernel run time Communication and arithmetic ranges Distributed Groups Distributed control where Process on SHPF F90 interface Distributed Distributed control Iterators on ranges Arrays Random ....

Bryan Carpenter, Yuh-Jye Chang, Georey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....sequential Fortran or C arrays, which can be passed to the standard MPI functions. Inquiry functions on distributed arrays return the sequential arrays as pointers or handles (depending on the language instantiation) We have already implemented a Java language binding for MPI, version 1. 1 [6, 8]. Our current approach is a relatively direct transcription of standard MPI bindings, but Java object serialization introduces new possibilities for passing compound objects. Similar projects on Java MPI bindings are in progress elsewhere [26, 15] 2.4 Integration of thread based single Java VM ....

Bryan Carpenter, Yuh-Jye Chang, Georey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....and describe an implementation using Java wrappers to invoke C MPI calls through the Java Native Interface [14] The software is publically available from http: www.npac.syr.edu projects pcrc mpiJava 1. 1 Related work Early work by two of the current authors on Java MPI bindings is reported in [3, 4]. In those papers we compared various approaches to parallel programming in Java, including socket programming and MPI programming. A comparable approach to creating full Java MPI interfaces has been taken by Getov and Mintchev [17, 13] In their work Java wrappers were automatically generated ....

Bryan Carpenter, Yuh-Jye Chang, Georey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....awkward. Rather than follow the HPF model directly, we propose introducing some of the characteristic ideas of HPF speci cally its distributed array model and array intrinsic functions and libraries into a basically SPMD programming model. Because the programming model is SPMD, direct calls to MPI [1] or other communication packages are allowed from the HPJava program. The language outlined here provides HPF like distributed arrays as language primitives, and new distributed control constructs to facilitate access to the local elements of these arrays. In the SPMD mold, the model allows ....

Bryan Carpenter, Yuh-Jye Chang, Georey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....awkward. Rather than follow the HPF model directly, we propose introducing some of the characteristic ideas of HPF speci cally its distributed array model and array intrinsic functions and libraries into a basically SPMD programming model. Because the programming model is SPMD, direct calls to MPI [2] or other communication packages are allowed from the HPJava program. The language outlined here provides HPF like distributed arrays as language primitives, and new distributed control constructs to facilitate access to the local elements of these arrays. In the SPMD mold, the model allows ....

....of b selected by the section. As in Fortran, upper or lower bounds can be omitted in triplets, defaulting to the actual bound of the parent array, and the stride entry of the triplet is optional. The subscripts of e, like any other array, start at 0, although the rst element is identi ed with a [2, 2]. In our language, unlike Fortran, it is not allowed to use vectors of integers as subscripts. The only sections recognized are regular sections de ned through scalar and triplet subscripts. The language provides a library of functions for manipulating its arrays, closely analogous to the array ....

Bryan Carpenter, Yuh-Jye Chang, Georey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....properties: 1. Components may be Java Beans or CORBA objects. They may be sequential codes written in Java, Fortran, or C; they may be parallel codes that make use of message passing libraries such as MPI; or they may exploit array based parallelism through language extensions such as HPJava [9]. Legacy codes, in Fortran for instance, can be wrapped as components. 2 2. Components themselves may be hierarchical (i.e. constructed from other components) and be of arbitrary granularity. Thus, a component may perform a simple task, such as nding the average of a set of input values, or it ....

B. Carpenter, Y.-J. Chang, G. Fox, D. Leskiw, and X. Li, \Experiments with HP Java," Concurrency: Practice and Experience, Vol. 9, no. 6, pp. 633-648, June 1997.

....support to explore the usage of Java in high performance parallel computing. Early experience on data parallel computing support, such as HPF [1] 7] language and its compiler technique research, showed that global arrays and or distributed arrays are the most powerful aspects in the languages [2], 6] In this article, we will focus on the potential of introducing some of the characteristic ideas of HPF specifically the distributed array model and array intrinsic functions and libraries as an extension to Java, into a basicly SPMD programming model. In this new data parallel ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....scattered, possibly non contiguously, over the local memory of a processor. These buffers are described through special objects called derived datatypes run time analogues of the user defined types supported by languages like C. The standard MPI approach does not map very naturally into Java. In [2, 3, 1] we suggested a Java compatible restriction of the general MPI derived datatype mechanism, in which all primitive elements of a message buffer have the same type, and they are selected from the elements of a onedimensional Java array passed as the buffer argument. This approach preserves some of ....

....MPI.BYTE byte MPI.CHAR char MPI.SHORT short MPI.BOOLEAN boolean MPI.INT int MPI.LONG long MPI.FLOAT float MPI.DOUBLE double MPI.OBJECT Object Figure 1: Basic datatypes in proposed Java binding 1. 2 Related work Early work by the current authors on Java MPI bindings is reported in [2]. A comparable approach to creating full Java MPI interfaces has been taken by Getov and Mintchev [14, 9] A subset of MPI is implemented in the DOGMA system for Java based parallel programming [11] A pure Java implementation of MPI built on top of JPVM has been described in [5] So far these ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HP- Java. Concurrency: Practice and Experience, 9(6):633, 1997.

....sequential Fortran or C arrays, which can be passed to the standard MPI functions. Inquiry functions on distributed arrays return the sequential arrays as pointers or handles (depending on the language instantiation) We have already implemented a Java language binding for MPI, version 1. 1 [6, 8]. Our current approach is a relatively direct transcription of standard MPI bindings, but Java object serialization introduces new possibilities for passing compound objects. Similar projects on Java MPI bindings are in progress elsewhere [26, 15] 2.4 Integration of thread based single Java VM ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....required for, eg, address translation. This can be seen as a new compiler targetting the run time an HPJava although the work of this compiler is much simpler than for HPF, because the programmer will explicitly specify the communication requirements. For early results of the work on Java see [20, 21, 46, 45, 50]. PCRC Java interface MPI Distributed data and control ad interface (Adlib) Kernel run time Communication and arithmetic ranges Distributed Groups Distributed control where Process on SHPF F90 interface Distributed Distributed control Iterators on ranges Arrays Random access schedules ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....and describe an implementation using Java wrappers to invoke C MPI calls through the Java Native Interface [14] The software is publically available from http: www.npac.syr.edu projects pcrc mpiJava 1. 1 Related work Early work by two of the current authors on Java MPI bindings is reported in [3, 4]. In those papers we compared various approaches to parallel programming in Java, including socket programming and MPI programming. A comparable approach to creating full Java MPI interfaces has been taken by Getov and Mintchev [17, 13] In their work Java wrappers were automatically generated ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....Rather than follow the HPF model directly, we propose introducing some of the characteristic ideas of HPF specifically its distributed array model and array intrinsic functions and libraries into a basically SPMD programming model. Because the programming model is SPMD, direct calls to MPI [1] or other communication packages are allowed from the HPJava program. The language outlined here provides HPF like distributed arrays as language primitives, and new distributed control constructs to facilitate access to the local elements of these arrays. In the SPMD mold, the model allows ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

....Rather than follow the HPF model directly, we propose introducing some of the characteristic ideas of HPF specifically its distributed array model and array intrinsic functions and libraries into a basically SPMD programming model. Because the programming model is SPMD, direct calls to MPI [2] or other communication packages are allowed from the HPJava program. The language outlined here provides HPF like distributed arrays as language primitives, and new distributed control constructs to facilitate access to the local elements of these arrays. In the SPMD mold, the model allows ....

....of b selected by the section. As in Fortran, upper or lower bounds can be omitted in triplets, defaulting to the actual bound of the parent array, and the stride entry of the triplet is optional. The subscripts of e, like any other array, start at 0, although the first element is identified with a [2, 2]. In our language, unlike Fortran, it is not allowed to use vectors of integers as subscripts. The only sections recognized are regular sections defined through scalar and triplet subscripts. The language provides a library of functions for manipulating its arrays, closely analogous to the array ....

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HPJava. Concurrency: Practice and Experience, 9(6):633, 1997.

No context found.

Bryan Carpenter, Yuh-Jye Chang, Geoffrey Fox, Donald Leskiw, and Xiaoming Li. Experiments with HP Java. Concurrency: Practice and Experience, 9(6), June 1997.

No context found.

B. Carpenter, Y.-J. Chang, G. Fox, D. Leskiw, and X. Li. Experiments with HP Java. Concurrency: Practice and Experience, 9(6), June 1997.

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