Page 85 - untitled

ҽ˱࿲

Lee, PeiZong ਿ ͉ ༟ ࣘ Research Description
༟
ࣘ
͉
ਿ
Research Description
My research interests are in compilers for The design of parallel architectures influences
ᔖcc၈j޼ӺࡰResearch Fellow (1998--)
scientific applications, parallel algorithm design, the design of parallel algorithms, and vice versa.
௰৷ኪዝj Ph.D., Computer Science, computer architectures, and the interplay among Compilers act as bridges connecting algorithms and
New York University (1989) architectures, algorithms, and compilers. The architectures. I am interested in studying this in-
design of parallel architectures represents cost- terdependence. I have studied various compilation
ཥcc༑j+886-2-2788-3799 ext. 1812 effective trade-offs. A general purpose machine issues on regular parallel architectures, focusing on
requires complex interconnection networks, multi- formal techniques, simulations, and the develop-
ෂccॆj+886-2-2782-4814
level memory hierarchies, powerful control units, ment of programming environments. The results
ཥɿඉ΁jleepe@iis.sinica.edu.tw and optimized schemes for allocation of machine obtained can be used both to compile algorithms on Research Fellows
resources. In contrast, to solve a class of problems such regular parallel architectures and to study the
ၣccࠫjhttp://www.iis.sinica.edu.tw/pages/leepe from an application domain or to share a common importance of various architectural features. Cur-
mathematical structure, it may be possible to design rent effort also has extended this approach to the
a special purpose machine with elementary inter- compilation of large scientific applications for mod-
connecting networks and elementary control units. ern general-purpose regular architectures.
޼Ӻᔊʧ
޼Ӻᔊʧ
Such a machine may also be able to solve this class
• PeiZong Lee received a B.S. degree in mathemat- of problems faster. I am currently also participating in a scientifi c
Research Fellows
i
ics from National Taiwan University in 1979, an ଣซʕd̻БஈଣᇜᙇኜᏐ༈ঐਗ਼ృҏό೻ό computation project, in which some colleagues and
M.S. degree in computer and decision sciences from Іਗᇜᙇϓ޴࿁̻ٙБஈଣཥ໘ٙੂБᇁdШ݊஢ In the context of parallel algorithms, sev- I have to conquer researches for unstructured mesh
generation, unstructured mesh partition, Euler equa-
National Tsing Hua University, Taiwan in 1984, and eral paradigms and techniques, such as divide and
εՍ৶თːٙ຾᜕Ⴍ׼ࠅ༺ՑІਗପ͛Ϟࣖଟٙੂ conquer, dynamic programming, domain decom- tion and Navier-Stokes equation solvers for a nu-
a Ph.D. degree in computer science from the Cou- merical wind tunnel platform, an engine combustion
Бᇁd֠ܙрɢfவԬࡡΪܼ̍jɓࡈϞࣖଟٙృ position, and program transformation, have been
rant Institute of Mathematical Sciences, New York studied. By using these paradigms, it is possible to platform for computing reactive fl ows, visualization,
University, in 1989. Between 1984 and 1986, he ҏόစၑجd̙ঐԨೌجٜટᔷ౬Ԩᇜᙇϓɓࡈ޴ design efficient parallel algorithms, at least in the- and the challenge of using MPI on workstation/PC
was a research engineer at the Institute of Informa- ࿁ήϞࣖଟ̻ٙБ೻όᇁfΪމ̀඲዆᜗Ͻᅇί዆ ory. Their practical utilization requires appropriate clusters to accelerate irregular computation. One of
tion Industry, Taiwan where he participated in an ࡈᏐ͜೻όʕ͟ක֐Їഐ҈΢ࡈԴ͜ՑٙစၑجՉ architectural and software support, techniques for my goals is to understand the techniques for compil-
ing irregular computation.
artificial intelligence project. He joined the Institute proving the correctness and efficiency of implemen-

ʱйٙ༟ࣘʱ௲ʿʱбd༟ࣘෂ৔dʿੂБર೻ٙ tation, and suitable programming environments.
of Information Science, Academia Sinica in 1989.
዆᜗௰Գʷf
His research interests are in compilers for scientifi c
applications, parallel algorithm design, computer
Selected Publications
࿁׵஝ۆٙᏐ͜೻όdͦۃʊ຾ঐ੄Іਗʱ Selected Publications
architectures, and the interplay among architectures,
algorithms, and compilers. ؓ዆ࡈᏐ͜೻όٙ༟ࣘݴ೻dԨ˲̙˸ପ͛Ϟࣖଟ 1. PeiZong Lee, Chien-Min Wang, and Jan-Jan Wu. “Compiler and Run- No. 2, Aug. 1995, pp. 72--85.

ٙੂБᇁfШ݊࿁׵ʔ஝ۆٙᏐ͜೻όd͟׵Դ͜ time Parallelization Techniques for Scientific Computations on Dis- 10. PeiZong Lee and Fang-Yu Huang. “An Effi cient Prime-Factor Algo-
tributed Memory Parallel Computers.” In book High Performance rithm for the Discrete Cosine Transform and Its Hardware Implemen-
ՑගટЗѧd֠͊Ϟ༰Ϟࣖଟٙ௰Գʷ༟ࣘʱؓҦ Computing: Paradigm and Infrastructure, Laurence T. Yang and tations.” IEEE Transactions on Signal Processing, Vol. 42, No. 8, Aug.
Minyi Guo, editor, pp. 135--181, John Wiley & Sons, Inc., 2006. 1994, pp. 1996--2005.
ஔfͦۃ༰̙Бٙ˙جܼ̍͟೻όႧԊ౤ԶʈЪʩ 2. PeiZong Lee, Chih-Hsueh Yang, and Jeng-Renn Yang. “Fast Algo- 11. PeiZong Lee and Fang-Yu Huang. “Restructured Recursive DCT and
rithms for Computing Self-Avoiding Walks and Mesh Intersections DST Algorithms.” IEEE Transactions on Signal Processing, Vol. 42,
ᄴܸࠦٙ˿ණdԨ˲͟೻όࡰ౤ԶᏐ͜ਪᕚᄴࠦٙ over Unstructured Meshes.” Advances in Engineering Software, Vol. No. 7, July 1994, pp. 1600--1609.
35, No. 2, Feb. 2004, pp. 61--73. 12. PeiZong Lee and Zvi M. Kedem. “On High-Speed Computing with a
ɽਜ෯ჯਹʱ௲စၑجdᇜᙇኜΎਗ਼ჯਹʱ௲စၑ 3. PeiZong Lee and Zvi M. Kedem. “Automatic Data and Computation Programmable Linear Array.” The Journal of Supercomputing, Vol. 4,
Decomposition on Distributed Memory Parallel Computers.” ACM No. 3, Sep. 1990, pp. 223--249.
جᇜᙇЇ̻БஈଣኜɪੂБf͟׵ྼ̻ٙ͜Бஈଣ Transactions on Programming Languages and Systems, Vol. 24, No. 1, 13. PeiZong Lee and Zvi M. Kedem. “Mapping Nested Loop Algorithms
Jan. 2002, pp. 1--50. into Multidimensional Systolic Arrays.” IEEE Transactions on Paral-
ཥ໘͊Ըٙᒈැ݊ϗⶊЇၣًࣸٙɽඎ̻Бஈଣኜ 4. PeiZong Lee and Wen-Yao Chen. “Generating Communication Sets of lel and Distributed Systems, Vol. 1, No. 1, Jan. 1990, pp. 64--76.
Array Assignment Statements for Block-Cyclic Distribution on Dis- 14. PeiZong Lee and Zvi M. Kedem. “Synthesizing Linear-Array Algo-
א͟ၣ༩ஹટٙᓉණεஈଣኜཥ໘ʈЪ१d͊Ը̻ tributed Memory Parallel Computers.” Parallel Computing, Vol. 28, rithms from Nested For Loop Algorithms.” IEEE Transactions on
No. 9, Sep. 2002, pp. 1329--1368. Computers: Special Issue on Parallel and Distributed Algorithms,
Бஈଣٙϓ̌ၾщਗ਼සӔ֛ί݊щঐ዆Υ̻Бစၑ
5. PeiZong Lee. “Efficient Algorithms for Data Distribution on Distrib- Vol. C--37, No. 12, Dec. 1988, pp. 1578--1598.

جʿᇜᙇኜЇϤՇɽᗳ̻Бஈଣཥ໘ɪf uted Memory Parallel Computers.” IEEE Transactions on Parallel 15. PeiZong Lee, Chih-Hao Chang, and Jan-Jan Wu. “Parallel Implicit
and Distributed Systems, Vol. 8, No. 8, Aug. 1997, pp. 825--839. Euler Solver on Homogeneous and Heterogeneous Computing Envi-

6. PeiZong Lee and Gau-Shin Liu. “An Efficient Algorithm for the 2-D ronments.” AIAA paper 2001-2588, 15th AIAA Computational Fluid
ί͊Ըٙ޼Ӻ˙ࠦdҢਗ਼ߧɢ׵೯࢝ɓࡈᅰ࠽ Discrete Cosine Transform.” Signal Processing, Vol. 55, No. 2, Dec. Dynamics Conference, Anaheim, California, June 11--14, 2001.
1996, pp. 221--239. 16. PeiZong Lee, Jan-Jan Wu, and Chih-Hao Chang. “Partitioning Un-
ࠬݸᅼᏝΌዚݴఙٙழ᜗ᐑྤ̨̻˸ʿɓࡈᅼᏝˏ 7. PeiZong Lee. “Techniques for Compiling Programs on Distributed structured Meshes for Homogeneous and Heterogeneous Parallel
Memory Multicomputers.” Parallel Computing, Vol. 21, No. 12, Computing Environments.” In Proc. the International Conference on
ᏗዷደˀᏐݴٙழ᜗ᐑྤ̨̻fҢҎૐঐዌ༆̻Б Dec. 1995, pp. 1895--1923. Parallel Processing, pp. 315--322, Vancouver, British Columbia, Au-
8. PeiZong Lee. “Parallel Matrix Multiplication Algorithms on Hyper- gust 18--21, 2002.
စၑجdᇜᙇኜd˸ʿɽۨʔ஝ۆ߅ኪࠇၑᏐ͜ਪ cube Multicomputers.” International Journal of High Speed Com- 17. PeiZong Lee, Chih-Hsueh Yang, and Jeng-Renn Yang. “Repartitioning
puting, Vol. 7, No. 3, Sep. 1995, pp. 391--406. Unstructured Meshes for the Parallel Solution of Engine Combustion.”
ᕚɧ٫ʘගٙʝਗᗫڷf 9. PeiZong Lee. “Mapping Nested-Loop Algorithms into Grid-Connected Presented at the Supercomputing'2003, Phoenix, AZ, Nov. 17-20,
Systolic Arrays without Data Collisions in the Data Links.” ISCA 2003.
International Journal of Computers and Their Applications, Vol. 2,

74 75