Iannucci, R. A., Gao, G. R., Robert H. Halsted, J. & Smith, B., eds [1994], Multithreaded computer architecture: A summary of the state of the art, Kluwer academic.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....independent operations. On the other hand, they are sequential architectures in the sense of the instructions because the processor has to check if an instruction is dependent of instructions in execution before issue it. Another example of hybrid architectures are the multithreaded processors [IGHS94] When executing the instructions inside a thread they are sequential architectures. At the thread level they are dependence architectures because there is a need to synchronize dependences between threads. 1.5 HARDWARE SUPPORT FOR EXTRACTING ILP Given a program with a particular static ....

R.A. Iannucci, G.R. Gao, R. H. Halstead, and B. Smith, editors. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer international series in engineering and computer science. Kluwer Academic Publishers, 1994.

....where code in high level languages, such as C, resembles instruction code much closer. In other words, we somewhat tried to resemble the selection of primitives in [Bl90] and Fork95 in our instruction set; other specific choices for the instruction set would have been possible. The edited book [IGHS94] is an excellent general reference for known approaches to multithreading. The approaches described there emphasize the case where in a given cycle instruction from only one thread can be issued. Another typical focus is on threads which are not too light as for instance in the compiler based ....

R.A. Iannucci, G.R. Gao, R.H. Halstead and B. Smith (Editors). Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer, Boston, MA, 1994.

....directory and by encouraging cache to cache transfers. It should be noted, however, that the directory access time is not changed in the memory latency experiments in Section 5.4. Another alternative is to use a more latency tolerant architecture, such as a multithreading at the instruction level [48]. ffl Average network latency is not a reliable indicator of application performance. In almost every experiment, invalidate is the worst performing mechanism, yet it has the lowest average network latency. StreamLine, on the other hand, often shows relatively high average latency, while ....

Robert A. Iannucci, Guang R. Gao, and Jr. Robert H. Halstead, editors. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer Academic Publishers, 1994.

....Cross Chip (Global) Communication Latency on XMT M Performance The U. Wisconsin Multiscalar project [13] and the U. Washington simultaneous multi threading (SMT) project [27] with their use of multiple program counters and the computer architecture literature on multi threading (see, for instance [17]) have also been very useful; however, the way XMT proposes to attack the completion time of a single task, which is so central to XMT, makes XMT drastically different than these approaches; that is, the reliance on PRAM algorithms. Our experience has been that some knowledge of PRAM algorithms is ....

R. A. Iannucci, G. R. Gao, R. H. Halstead, and B. Smith (editors). Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer, Boston, MA. 1994.

....block the parent process until the computation is nished. The incorporation of conventional control AEow thread execution into the dataAEow approach resulted in the multithreaded computer architecture which is one of the most promising and exciting avenues for the exploitation of parallelism [10,43,87,128,151,159]. 2 The day before yesterday: Pure dataAEow The fundamental principles of dataAEow were developed by Jack Dennis [65] in the early 1970s. The dataAEow model [8,68,93,191] avoids the two features of von Neumann model, the program counter and the global updatable store, which become bottlenecks in ....

R.A. Iannucci, G.R. Gao, R. Halstead, and B. Smith, Multithreaded computer architecture: A summary of the state of the art, Kluwer Academic, 1994.

....contains program codes and FM(Frame Memory) contains frames required to execute programs. Global data are stored in SM(Structured Memory) which is inside NIMU. None of nodes can directly load or store data into SM, i.e. global data are always remote and can be accessed through split transaction[13]. Figure 1: Node Architecture of DAVRID TPU executes threads in a sequential manner. The information of a thread to be executed is represented as fp, tp (fp frame base address, tp thread starting address) called continuation and fetched from ATQ(Active Thread Queue) whenever a thread ....

R. A. Iannucci, Multithreaded Computer Architecture: A Summary of the State of the Art, Kluwer Academic Publishers, 1994.

....latency for many years. Multithreaded processors have been proposed or used in several machines, such as the HEP [31] Monsoon [28] the Tera MTA [5] McGill MTA [19] and MIT Alewife [3] A description of the history and status of multithreaded architectures and research appears in [9] and [20]. Early multithreaded architectures take an aggressive fine grain approach that switches contexts at each instruction. This requires a large number of contexts and very low overhead context switches. The main drawback of this approach is poor single thread performance. To address this ....

R. A. Iannucci, editor. Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer Academic Publishers, 1994.

....Background Multithreading and prefetching have been the focus of work on tolerating communication latency for many years. Multithreaded processors have been proposed or used in several machines, such as the HEP [28] Monsoon [25] the Tera MTA [4] McGill MTA [16] and MIT Alewife [1] 7] and [17] describe the history and status of multithreaded architectures and research. Most of the earlier multithreaded architectures follow an aggressive fine grain approach to multithreading, where a large number of contexts are available and a context switch takes a small number of cycles. The main ....

Robert A. Iannucci, editor. Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer Academic Publishers, 1994.

....of achieving very high processor throughput. The introduction of the data structure cache reduces the network latency significantly. The impact of various cache organizations on the performance of the architecture is also discussed in this paper. 1 Introduction Multithreaded architectures [10, 11] are based on a hybrid evaluation model which combines the von Neumann execution model and the data driven evaluation. In the hybrid model, a program is represented as a partially ordered graph of nodes. The nodes, called threads, consist of a sequence of instructions which are executed in the ....

....approaches like PLAs or 3 A study in [14] shows that on an average one read and one read write port is sufficient for one instruction execution. FPGAs. 6 Related Work Several multithreaded architectures have been proposed in the literature (refer, for example, to [2, 3, 6, 12, 13, 15] and to [11] for a survey) Like Threaded Abstract Machine (TAM) 6] and T [15] our architecture realizes three levels of program hierarchy based on synchronization and scheduling. TAM uses a compiler controlled approach to achieve fine grain parallelism and synchronization. In contrast, we advocate the use ....

R. A. Iannucci, G. R. Gao, R. H. Halstead, Jr., and B. Smith. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer, Norwell, Mass., 1994.

....the throughput. The impact of various cache organizations on the performance of the architecture is also discussed in the paper. 1 Introduction Multithreaded architectures are based on a hybrid evaluation model which combines the von Neumann execution model and the data driven evaluation [10, 11]. In the hybrid model, a program is represented as a partially ordered graph of nodes. The nodes, called threads, consist of a sequence of instructions which are executed in the conventional von Neumann way. Individual threads are scheduled in a dataflow like manner, driven by the availability of ....

....activations Logic in these units is fairly very small and they can be implemented as finite state machines, using conventional approaches like PLAs, FPGAs. 6 Related Work Several multithreaded architectures have been proposed in the literature (refer, for example, to [2, 3, 6, 12, 13, 15] and to [11] for a survey) Like Threaded Abstract Machine (TAM) 6] and T [15] our architecture realizes three levels of program hierarchy based on synchronization and scheduling. TAM uses a compiler controlled approach to achieve fine grain parallelism and synchronization. In contrast, we advocate the use ....

R. A. Iannucci, G. R. Gao, R. H. Halstead, Jr., and B. Smith. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer, Norwell, Mass., 1994.

....latency for many years. Multithreaded processors have been proposed or used in several machines, such as the HEP [25] Monsoon [22] the Tera MTA [4] McGill MTA [14] and MIT Alewife [2] A description of the history and status of multithreaded architectures and research appears in [7] and [15]. Early multithreaded architectures take an aggressive fine grain approach that switches contexts at each instruction. This requires a large number of contexts and very low overhead context switches. The main drawback of this approach is poor single thread performance. To address this limitation, ....

R.A. Iannucci, editor. Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer Academic Publishers, 1994.

....only communicate to its immediate neighbor: long distance communication is mostly limited to servicing memory requests; even the address resolution mechanism is decentralized [11] The XMT architecture differs in that hardware contexts never communicate directly with each other. The edited book [17] is an excellent general reference for known approaches to multithreading. The approaches described there emphasize the case where in a given cycle instructions from only one thread can be issued. A very interesting architecture is presented in the Tera computer [1] which linkedlist (5000 item ....

R.A. Iannucci, G.R. Gao, R. H. Halstead, and B. Smith (editors). Multithreaded Computer Architecture - A Summary of the State of the Art. Kluwer, Boston, MA. 1994.

....called latency tolerant architectures. 1 A history of the evolution of the dataflow computation model(s) and the variety of architecture and computational models that have been inspired by it is beyond the scope of this paper. The reader is referred to, among many others, the following texts [63, 55, 41]. As a computation model, the dataflow approach has had influence on many areas of computer science and engineering research. Examples include programming languages, processor design, multithreaded architectures, parallel compilation, high level logic design, signal processing, distributed ....

....dataflow or dynamic dataflow. A number of articles have been published on multi threaded execution and architecture models with dataflow origin and can found in a survey article by Dennis and Gao [29] Principal projects and representative work before 1995 have been discussed in two monographs [55, 41] as well as other survey articles [61] Below we briefly mention a few research projects on each model. Inspired by the static dataflow model, the McGill Dataflow Architecture Model [39] has been proposed based on the argument fetching principle [28] The architecture departs from a direct ....

[Article contains additional citation context not shown here]

R. A. Iannucci, G. R. Gao, R. H. Halstead, Jr., and B. Smith, editors. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer Academic Publishers, Norwell, Massachusetts, 19 94. Book contains papers presented at the Workshop on Mult ithreaded Computers, Albuquerque, New Mexico, November 1991.

....we give our first quantitative indications of how much the basic multithreading support can help in tolerating increasing communication synchronization demands. Keywords: multithreaded architectures, multiprocessors, performance measurements, EARTH. 1 Introduction Multithreaded architectures [1, 2, 3, 4, 5, 14, 15, 16, 18] have been promoted as potential processing nodes for future parallel systems due to their toleration of inherent parallel In the Proceedings of the International Conference on Parallel Architectures and Compilation Techniques (PACT 95) Limassol, Cyprus, June 27 29, 1995, pages 59 68. Copyright ....

Robert A. Iannucci, Guang R. Gao, Robert H. Halstead, Jr., and Burton Smith, editors. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer Academic Pub., Norwell, Mass., 1994.

No context found.

Iannucci, R. A., Gao, G. R., Robert H. Halsted, J. & Smith, B., eds [1994], Multithreaded computer architecture: A summary of the state of the art, Kluwer academic.

No context found.

R.A. Iannucci, editor. Multithreaded Computer Architecture: A Summary of the State of the Art. Kluwer Academic Publishers, 1994.

No context found.

R. Iannucci, ed., Multithreaded Computer Architecture: A Summary of the State of the Art, Kluwer Academic Publishers, 1994.

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