|
|
 |
|
|
Distributed Parallel Processing Laboratory |
Nikolay N. Mirenkov
Professor |
Stanislav G. Sedukhin
Professor |
Oleg Monakhov
Visiting Professor |
Rentaro Yoshioka
Visiting Professor |
|
Our general program is cyber infrastructure and high-performance computing. We are also thinking about a peer-to-peer and grid computing, as well as about active knowledge being developed by humanity. We are undertaking research efforts in parallel and distributed algorithms/architectures, visual (multimedia) languages and tools.
We are studying the problem of parallel program portability, skeleton-based techniques and developing the supporting tools. Our theoretical work is related to efficient algorithms/architecture for highly-parallel numerical computation including linear algebra, multimedia, and network processing.
We have proposed a new scalable parallel processor architecture for the future VLSI technology. This processor (code name Trident) is able to support the scalar, vector, and matrix instructions and based on local communications not only within but also between processing elements. We have shown that proposed architecture can be effectively used for many scientific, engineering, and multimedia applications since they are mostly based on mixing of scalar, vector, and matrix parts.
The important question of our research is parallel program transparency and accessibility of massively parallel computers to large user populations. We are developing amultimedia technology for the interactive specification of application algorithms. This technology is based on self-explanatory components in a cyberFilm format. A cyberFilm is a series of color stills supported, if necessary, bytext and voice/sound. Each still is to represent a view (some features) of objects or processes. Each cyberFilm is to represent a multiple view (an extended set of dynamic and/or static features) of objects or processes. Different views can be related to different moments of time, positions in space, levels of hierarchy, values of data attached to space points, etc. Different media can be used for different views. A self-explanatory cyberFilm means that the associated stills are organized and presented in such away that the semantic richness is clearly brought out. The investment of meaning in the cyberFilm is reduced to developing a series of views watching (and hearing) in non-linear order. Usually, a still is self-evident and a cyberFilm is a result of special gathering of clues or hints. This result is a piece of knowledge. So, self-explanatory adequacy depends on this knowledge. The more accurate and relevant views are used, the greater adequacy is reached.
A computational method cyberFilm is a series of stills displaying one or more parameterized sets of nodes and/or moving objects in multi-dimensional spacetime. Each still highlights a subset of these nodes and/or moving objects. Each cyberFilm defines a partial order of scanning of the nodes or objects. As a rule, computation specified on different nodes (objects) of a still is considered to be performed in parallel. Computation specified in different stills is performed sequentially. So, it is possible to say: the shorter ??lm the better. The user defines the specification by creating his new film. The corresponding program (sequential or parallel) is generated automatically. In fact, we are developing cyberFilm machines where data, knowledge and algorithms (as well as results) are specified by cyberFilms (self-explanatory components). We lead two projects related to filmification of methods and data: 1) Active Knowledge Studio for teachers, students, and programmers, and 2) F-Communication System for children, elderly and handicapped people,
|
| [nikmir-01:2002] | N. Mirenkov R.Yoshioka. Visual Computing within Environment of Self-explanatory Components. Soft Computing Journal, 14(3):215-232, 2002.
A visual language and amultimedia environment supporting the language are considered. The language is explained through describing a visual program for solving partial differential equations by amultigrid method. The environment is based on a database of self-explanatory components in a film format and a film management system for searching, editing, composing and other manipulations with components. The visual program presented is only one view of a self-explanatory component. In fact, it can also be watched in dynamics and from other points of view for better understanding the method features. The goal of the visual program example is to show that visual programs can be physically much smaller than the text they replace and much easier for understanding and modifying. |
| [nikmir-02:2002] | Tetsuya Hirotomi and Nikolay Mirenkov. Self-explanatory components: a basis for new communicators. Journal of Visual Languages and Computing, 1(14):215-232, 2003.
An approach for using the self-explanatory component concept in multimedia message representation is considered. It is based on cyberFilm formats of multimedia symbols (multimedia words) and multimedia sentences (multimedia hieroglyphs). Corresponding films are multiple views of objects, processes, etc. They are pieces of knowledge. These pieces are acquired in a film database. The user should not study them in advance. A film management system (including self-explanatory interface panels) provides effective access to the database items. In this paper, the concept of self-explanatory components, formats to represent multimedia sentences and words, as well as a brief explanation of the film (component) management system are provided. Some implementation results oriented to people with communication disorders are presented.are provided. |
| [nikmir-03:2002] | Rentaro Yoshioka Yutaka Watanobe and Nikolay Mirenkov. Self-explanatory components to study algorithms. Journal of Three Dimensional Images, 16(4):215-232, 2002.
A new approach using self-explanatory components for education is considered. This approach is based on a cyberFilm format of algorithms and on a cyberFilm management system. A brief description of self-explanatory components and visual programming environment is provided. In addition, examples of cyberFilm representation, sorting algorithms and their multiple views are presented. |
| [rentaro-01:2002] | Y. Watanobe, R. Yoshioka, and N. Mirenkov. Self-explanatory Components to Study Algorithms. Journal of Three Dimensional Images, 16(4):231-236, 2002.
We present a new approach for education of algorithms using self-explanatory components. |
| Referred Proceeding Papers |
| [nikmir-04:2002] | Robert R. Roxas and Nikolay N. Mirenkov. Visualizing External Inter-Component Interfaces. In Proc. of the 22nd IEEE Int. Conference on Distributed Computing Systems, Workshops, pages 290-295, Vienna, July 2002.
Despite the advances in software technology, there are still problems that need to be solved. They are related to understanding components, modifying them and their interfaces, etc., as well as to the practical use of Component-Based Development (CBD) for large-scale applications. Our approach is basedondevelopingmultiple view components of multimedia types. These multiple views can be divided into a few groups to represent different features of acomponent: (1) computational schemes of a corresponding component algorithm, (2) variables and formulas used in the algorithm, and (3) input/output (I/O) operations of the component. Such approach simplifies the understanding and manipulation of components. This paper discusses examples of visualizing I/O specifications which include the definition of data source and the corresponding target, scanning schemes applied to structures, conditions imposed upon different operations, etc .In other words, a visual language to define external intercomponent interfaces is presented. |
| [nikmir-05:2002] | Robert R. Roxas and Nikolay N. Mirenkov. Visualizing Input/Output Speci??cation. In The Eighth Int. Conference on Distributed Multimedia Systems, pages660{663, SanFrancisco, Sep.2002.
In software components, there are several kinds of input/output (I/O) communication to the outside world. This I/Ocommunication includes the conventional type of I/O operations such as using the keyboard, mouse, and Graphical User Interfaces (GUIs). Other types of I/O communication include the communication between processes in parallel programs, between client stubs and server stubs, between twocomponents in a complex system, and between a component and a file. Reusability of software components depends, in a great part, on how can we understand and update the component interfaces. This paper provides a brief overview of our approach to visualizing computational algorithms and presents one example of input operation specification for inter-component communication. A visual language of micro-icons for specifying I/O operations between components and a special panel supporting this specification are considered. |
| [nikmir-06:2002] | Mirenkov N. Tsuchida Y. Yoshioka, R. and Y. Watanobe. Visual Notation of Film Language System. In Proc., 2002 International Conference on Distributed Multimedia Systems (DMS-2002), pages 648-655, San Francisco, Sep. 2002.
The Film language system is a software framework to specify, present and explain computational algorithms through their multiple views collected in a 'film' format. This system includes a film database and a film management system to support the acquiring, rendering, editing and creating of films. The film format is a generalization of a conventional movie concept that introduces new types of frames and a non-linearon-demand access to frames. This paper presents an overview of notations used in the film language to represent frames, scenes, and lims. |
| [nikmir-07:2002] | M. Saber and N. Mirenkov. Visualizing cellular image processing algorithms. In Proc., 2002 International Conference on Distributed Multimedia Systems (DMS-2002), pages 656-659, San Francisco, 2002.
The paper enlightens the idea of using cellular automatain image processing, presents a new visual representation for cellular image processing algorithms as one of the implementations of self-explanatory component technology, where self-explanatory features are reached by special collections of component views. |
| [nikmir-08:2002] | T. Hirotomi and N. Mirenkov. Multiple interface communication system. In Proc., 2002 International Conference on Distributed Multimedia Systems (DMS-2002), pages 606{611, San Francisco, 2002.
Amultimedia communication environment for people with different levels of disorders is presented. A brief explanation of multimedia messages, sentences and words is provided and the system architecture and multiple interfaces are described. |
| [nikmir-09:2002] | Monakhov O. Mirenkov, N. and R. Yoshioka. Self-explanatory components: visualization of graph algorithms. In Proc., 2002 International Conference on Distributed Multimedia Systems (DMS-2002), pages 562-567, San Francisco, 2002.
A novel approach for using the self-explanatory component concept in the graph algorithm representation is considered. It is based on a film format, where a series of multimedia frames represents a set of algorithm features. A brief description of the film format is provided, and an observation of graph algorithm features is presented. An example related to Prims algorithm is used to explain the basic ideas of the approach. Two versions of the algorithm visualization, as well as film frames and an icon language of the frames are demonstrated. |
| [nikmir-10:2002] | Monakhov O. Mirenkov, N. and R. Yoshioka. Visualization of graph algorithms and programming in pictures. In Proc. of the 6th IASTED Con. on Software Engineering and Applications, pages 391-397, Cambridge, USA, 2002.
A new approach for using pictures, icons and animation to specify, present, explain, and programming algorithms is considered. The consideration is based on a number of examples related to Prim and Dijkstra algorithms. Several versions of these algorithm visualization, as well as corresponding pictures, icons and animation, are demonstrated and/or explained. |
| [nikmir-11:2002] | Yoshioka R. Ebihara, T. and N. Mirenkov. Program generation from film specifications. In Proc. of the 6th IASTEDCon. on Software Engineering and Applications, pages 403-410, Cambridge, USA, 2002.
An approach to program generation from the film specification is presented. An example of template programs supporting the approach is also provided. |
| [nikmir-12:2002] | T. Hirotomi and N. Mirenkov. Multimedia Communication Environment for Children, Handicapped, and Elderly People. In Proceedings of the 22nd International Conference on Distributed Computing Systems, Workshops, pages 122-127, Vienna, 2002.
Self-explanatory components for the communication of people with special needs and an overview of two interfaces of the corresponding system are presented. |
| [nikmir-13:2002] | R. Roxas and N. Mirenkov. Visual Input/Output Specification in Different Modes. In Proc. of the 17th International Conference on Advanced Information Networking and Applications, pages 185-188, Xian, China, 2003.
Watching, editing and composing modes to perform manipulations with frames of cyberFilms representing Input/Output speci??cation are considered. |
| [rentaro-02:2002] | N. Mirenkov, O. Monakhov, and R. Yoshioka. Visualization of Graph Algorithms and Programming in Pictures. In Editor M.H. Hamza, editor, The Sixth IASTED International Conference on Software Engineering and Applications (SEA?2002), pages 391-397, Cambridge, Nov. 2002. IASTED, ACTA Press.
We present our approach in using pictures, icons and animations to specify, explain, and program graph algorithms. |
| [rentaro-03:2002] | T. Ebihara, R. Yoshioka, and N. Mirenkov. Program Generation from Film Specification. In Editor M.H. Hamza, editor, The Sixth IASTED International Conference on Software Engineering and Applications (SEA?2002), pages 403-410, Cambridge, Nov. 2002. IASTED, ACTA Press.
We present our approach in generating programs from a ??lm specification, which is a new type of abstraction for representing multiple views of algorithms. |
| [rentaro-04:2002] | N. Mirenkov, R. Yoshioka, Y. Tsuchida, and Y. Watanobe. Visual Notation of FilmLanguage System. In Editor S.K.Chang, editor, The Eighth International Conference on Distributed Multimedia Systems (DMS?2002), pages 648{655, California, Sept. 2002. Distributed Multimedia Systems, Knowledge Systems Institute.
We propose a system of visual languages to represent multiple views of algorithms. |
| [rentaro-05:2002] | N. Mirenkov, O. Monakhov, and R. Yoshioka. Self-explanatory Components: Visualization of Graph Algorithms. In Editor S.K. Chang, editor, The Eighth International Conference on Distributed Multimedia Systems (DMS?2002), pages 562-567, California, Sept. 2002. Distributed Multimedia Systems, Knowledge Systems Institute.
We describe our approach in representing computational algorithms related to graph structures. |
| [sedukhin-01:2002] | M. Soliman and S. Sedukhin. BLAS on the Trident Processor: Implementation and Performance Evaluation. In Editor N. Debnath, editor, 18th International Conference on Computers and their Applications (CATA-2003), pages 359{364, Cary, NC, March 2003. ISCA, ISCA.
This paper describes the implementation of the Basic Linear Algebra Subprograms (BLAS), which are widely used in many applications, on the Trident processor. We show how to use the Trident parallel execution units, ring, and communication registers to effectively perform vector-vector, matrix-vector, and matrix-matrix operations needed for implementing BLAS. The TFLOPS rate on infinite-size problems (Rinf), which is primarily a characteristic of the computer technology, and the problem size needed to reach one-half of Rinf (N1/2), which is a measure of the amount of parallelism in a computer architecture, are used to evaluate the performance of the Trident processor on BLAS. On 128 parallel Trident lanes and 10 GHz clock frequency, which are possible in the billion-transistor era, Rinf of dot-product, matrix-vector, and matrix-matrix multiplications are 1.1, 1.8, and 2.5 TFLOPS, respectively. Besides, N1/2 increases when switching from low level to high level of BLAS. |
| [sedukhin-02:2002] | M. Soliman and S. Sedukhin. Performance Analysis of SVD Algorithm on the Trident Processor. In S. Yukita Editors S. Peng, V. Savchenko, editor, The First International Symposium on Cyber Worlds: Theory and Practices (CW 2002), pages 95{102, New York, Nov. 2002. IEEE, IEEE Computer Society Press.
In this paper, the performance of singular value decomposition (SVD) algorithm is evaluated on the Trident processor. We show that Trident ISA significantly reduces the number of instructions, loop overhead, and load/store operations needed for SVD. Moreover, Trident processor is scalable and its scalability needs only replicating lanes to process longer vectors or larger matrices. |
| [sedukhin-03:2002] | S. Sedukhin and S. Peng. Design of Array Processors for Multidimensional Image/Signal Processing. In Editor M. Sambandham, editor, The Second International Conference on Neural, Parallel, and Scientific Computations, pages 113-118, Atlanta, USA, Aug. 2002. Morehouse College, Dynamic Publ.
In this paper, a new method for designing effcient array processors for multidimensional image/signal processing is proposed. The major advantage of this method is that the design of effcient array processors is reduced to the design of effcient linear array processors which are much easier to handle than the multi-dimensional ones. Another advantage of this method is that no transpo- sition of any intermediate matrix is needed through whole transformation. The computations of different dimensions can be done consecutively without additional data movements in between. This method is particularly attractive for multi-dimensional image processing which is essential in recent digital video and multimedia applications. Based on the proposed design method we propose new arrayprocessors for computing multi-dimensional image transforms including DFT (Discrete Fourier Transform), DCT (Discrete Co- sine Transform), and OHT (Ordered Hadamard Transform). The design of the array processors for the r-D image transforms depends completely on the design of corresponding linear array processors for the 1-D case. Therefore, we focus on designing I/O and communication effcient schemes for 1-D DFT, DCT, and OHT. Our schemes eliminate the need for generating the whole kernel matrix of the 1-D image transforms. The coeffcients are created locally on the y when- ever needed. These I/O effcient recursive schemes are of particular interests when the real-time systems such as digital video are considered. The proposed array processors for the r-D image transforms are scalable, regular, and fully-pipelined. Therefore, they are suitable for high speed video/multimedia image/signal transmission in real time. |
| [sedukhin-04:2002] | M. Soliman and S. Sedukhin. Multi-level ISA Processor for Accelerating Data Parallel Applications. In Editor H.R. Arabnia, editor, International Conference on Parallel and Distributed Techniques and Applications (PDPTA'2002), pages 1492-1498, Las Vegas, NV, March 2002. CSREA, CSREA Press.
Different subtasks of an application usually have different computational and I/O requirements that result in different needs for processor capabilities. Thus, the more appropriate approach for both high performance computing and simple programming model is designing a processor having multi-level instruction set architecture (ISA). Each level has instructions executed on a different data structure and precisely telling the processor what the pro-ram needs done in compact form, leading to high performance and minimum executable code size. Since the fundamental data structures for a wide variety of data parallel applications are scalar, vector, and matrix, our proposed Trident processor has three-level ISA executed on zero-, one-, and two-dimensional arrays of data. These levels are used to express a great amount of fine-grain parallelism (up to three levels of data parallelism) to a processor instead of using complex hardware techniques to dynamically extract parallelism. This reduces the design complexity and provides high-level programming interface to hardware. The Trident processor consists of a set of parallel vector pipelines (PVPs) combined with a scalar core. The PVPscan access ring registers to perform vector and matrix operations. The scalability of the Trident processor does not require more fetch, decode, or issue bandwidth, but requires replication of PVPs and increasing the register file size. Scientific, engineering, multimedia, neural network, and many other applications, which are based ona mixture of scalar, vector and matrix operations, can be eAEciently speeded up on the Trident processor. |
| [nikmir-14:2002] | N. Mirenkov. Ministry of Education Scientific Research Fund, 2001-2003. |
| [nikmir-15:2002] | N.Mirenkov. Knowledge Cluster Creation Support Project, 2002-2003. |
| [nikmir-16:2002] | N. Mirenkov. Fukushima Prefectural Foundation, 2002. |
| [nikmir-17:2002] | Nikolay Mirenkov, Sep. 2002. Co-chair of the Program Committee, International conference DMS-2002, USA |
| [nikmir-18:2002] | Nikolay Mirenkov, 2002. Member, the IFIP Working Group 10.3 (Concurrent Systems) |
| [nikmir-19:2002] | Nikolay Mirenkov, 2002. Associate Editor, the Tamkang Journal of Science and Engineering, International Journal |
| [nikmir-20:2002] | Nikolay Mirenkov, 2002. Member, ACM and IEEE |
| [sedukhin-05:2002] | S Sedukhin, Apr. 2002. IEEE CS, member |
| [sedukhin-06:2002] | S Sedukhin, Apr. 2002. ACM,member |
| [sedukhin-07:2002] | S Sedukhin, Apr. 2002. IASTED Technical Committee on Parallel Processing, member |
| [sedukhin-08:2002] | S Sedukhin, Apr. 2002. Associative Editor for the IEICE Trans., Special Issue |
| [sedukhin-09:2002] | S Sedukhin, Apr. 2002. International Journal `Neural, Parallel & Scienti??c Computations', Member of the Editorial Board |
| [sedukhin-10:2002] | S Sedukhin, Apr. 2002. International Journal `Parallel Processing Letters', Member of the Editorial Board |
| [sedukhin-11:2002] | S Sedukhin, Nov. 2002. International Symposium on Cyber Worlds: Theory and Practice (CW2002),Program Committee Co-Chair |
| [sedukhin-12:2002] | S Sedukhin, Feb. 2003. 21st IASTED International Multi-Conference APPLIED INFORMATICS (AI 2003),Program Committee Member |
| [sedukhin-13:2002] | S Sedukhin, June 2002. 16th Annual International Symposium on High Performance Computing Systems and Applications (HPCS 2002), Program Committee Member |
| [sedukhin-14:2002] | S Sedukhin, Sep. 2002. 3rd International Conference on Parallel and Distributed Computing, Applications, and Technologies (PDCAT 2002), Program Committee Member |
| [sedukhin-15:2002] | S Sedukhin, Aug. 2002. 2nd International Conference on Neural, Parallel, and Scientific Computations, Program Committee Member |
| [nikmir-21:2002] | Masaaki Nemoto. Graduation Thesis: Program Generator within Multimedia Programming Environment, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-22:2002] | Rikio Nomoto. Graduation Thesis: Filmification of Methods: Thermal Conduction Problem with the Gauss-Jordan Method, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-23:2002] | Ayako Oomiya. Graduation Thesis: Multimedia World Dictionaries for Vocabulary Development, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-24:2002] | Etsuko Nakamura. Graduation Thesis: Hieroglyph Literacy Education with Teachers, Robot-toys, and Screen-pets, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-25:2002] | Yukitake Imagawa. Graduation Thesis: A Multimedia Communicator Based on Self-explanatory Components, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-26:2002] | Masafumi Yamada. Graduation Thesis: Development of a Multimedia Communicator for Blind People, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-27:2002] | Kazuki Fujita. Master Thesis: Picture-based Interface Panels, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-28:2002] | Sakiko Matsuda. Master Thesis: Glossary Type Help System, University of Aizu, 2002.
Thesis Advisor: Nikolay Mirenkov |
| [nikmir-29:2002] | Tetsuya Hirotomi. PhD Thesis: Multiple Interface Multimedia Communication System, University of Aizu, 2003.
Thesis Advisor: Nikolay Mirenkov |
| [sedukhin-16:2002] | Toshio Saze. Graduation Thesis: SSE Implimentation of Level 2 BLAS on Pentium III Processor, University of Aizu, 2002.
Thesis Advisor: Sedukhin, S. |
| [sedukhin-17:2002] | Nobuyuki Adachi. Graduation Thesis: Matrix Multiplication Using Multi-level of BLAS on Pentium III Processor, University of Aizu, 2002.
Thesis Advisor: Sedukhin, S. |
| [sedukhin-18:2002] | Shinya Meguro. Master Thesis: Real-time communication using wireless LAN, University of Aizu, 2002.
Thesis Advisor: Sedukhin, S. |
| [sedukhin-19:2002] | Christina Belograzova. Master Thesis: Hard Real-Time Operating System for Cluster of Computers, University of Aizu, 2002.
Thesis Advisor: Sedukhin, S. |
|
Annual Review 2002