centers will charge a penalty for single users who want to grab all four processors. Yoshihara also discussed some aspects of this in benchmark calculations earlier this year [H. Yoshihara, "Performance of Japanese supercomputers vis-a-vis Cray computers," Scientific Information Bulletin 15(3), 65-74 (1990)]. At least three or four uniprocessor systems have been sold, in Europe. We were not told about sales of two- or four-processor systems. Users can write Fortran without any special directives. NEC provides an automatic parallelizing and vectorizing compiler option. We had no opportunity to test this. Watanabe showed us results of running 100 by 100 LINPACK (all Fortran), giving performance on the SX-3 model 13 (uniprocessor) and several other supercomputers as shown in Table 1. He also showed some corresponding figures for 1000 by 1000 linear system and for 1024 by 1024 and 4096 by 4096 matrix multiplication. The last two columns correspond to what Dongarra calls "best effort." There are no restrictions on the method used or its implementation. Matrix multiplication runs almost at theoretical peak speed. The large linear system runs at slightly less than 70% of peak, while on the Cray the same calculation runs at just above 80%. The differences are probably associated with bandwidth from memory to the vector registers. Nevertheless, at 3.8 GFLOPS the SX-3 is 80% faster than the Cray. To the best of our knowledge, figures for the NEC and Fujitsu machines are new. We asked Watanabe if the SX-3 four-processor performance would scale up, and he only exclaimed "God knows." NEC's chip technology is very good. Using ECL, they have crammed 20,000 gates with 70 picosecond switching time onto one chip. We think that this is better than in the United States. A 1,200-pin multichip package can hold 100 such chips and dissipate 3K watts. Packaging, carrier, and cooling technology is about as good as in the United States. NEC claims that they have taken extra care to design in error testing capability and that about 30% of their chip area is associated with diagnostic functions. (This is certainly different from some U.S. manufacturers.) The memory system uses 20-ns, 256-Kbit SRAMS. A memory card can hold 32 MB. Thus a memory cabinet with 32 memory cards has 1 GB. Two peripherals are worth noting. NEC makes a cartridge tape unit (IBM-compatible tapes), fully automated, with 1.2 terabyte capacity. NEC also makes a disk array made of eight byte-interleaved disks. Used as a single disk drive, the disk array has a 5.5-GB capacity. The burst transfer rate is 19.6 MB/s, whereas the sustained transfer rate is 15.7 MB/s. NEC has begun publication of a newsletter about the SX-3, SX World. Interested readers can obtain a copy by writing NEC, 1st Product Planning Department, EDP Product Planning Division, 7-1 Shiba 5-chome, Minato-ku, Tokyo 108-01, Japan. In this their view of supercomputing is stated explicitly dVP2600 model was not specified for the Ax-b figures and was /10 for A=B*C, but both 2600/10 and /20 have the same peak performance, 5 GFLOPS. e4096 x 4096 The actual performance of a supercomputer is determined by its scalar performance ... NEC's approach to supercomputer architecture is clear. Our first priority is to provide high-speed single processor systems which have vector processing functions and are driven by the fastest technologies, while giving due consideration to ease of programming and ease of use; we also seek to provide shared memory multiprocessor systems to further improve performance. The SX-3 looks like an exciting machine that is on a par with the best currently available U.S. products. There is a new U.S. supercomputer from Cray Research nearly ready to be released, as well as perhaps models from Cray Computer Corporation and others, but we have no concrete information about their performance. In its four-processor version, the SX-3 might be the fastest large-scale supercomputer, but this will be entirely dependent on the application and the skill of the compiler writers. Fujii and Tamura (“Capability of current supercomputers for computational fluid dynamics," Institute of Space and Astronautical Science, Yoshinodai 3-1-3, Sagamihara, Kanagawa 229, Japan) note that "basically the speed of the computations simply depend on when the machines were introduced into the market. Newer machines show better performance, and companies selling older machines are going to introduce new machines."--David K Kahaner, ONRASIA KANAGAWA SCIENCE PARK Kanagawa Science Park (KSP) is an urban science park in Kawasaki. It was completed a year ago and is the largest of eight research core parks built or being planned under the Private Sector Resources Utilization Law. The eight are KSP, Eniwa Research Business Park (Hokkaido), 21st Century Plaza (Sendai), Tsukuba Research Support Center, Kurume Techno Research Park (Kyushu), Senri Life Science Center (Osaka), Toyama Advanced Industry Base, and Nagaoka Research Core. Of these, the last three are not yet complete. These research parks provide research and development (R&D), training, conference, and management facilities to promote R&D and new innovative industries in various regions of Japan. KSP is a very impressive modern facility consisting of two large buildings. It was built for ¥65B ($500M at ¥130-$1). The funds were provided by the Governments of Japan, Kanagawa Prefecture, and Kawasaki. The Japanese Government monies came from the Ministry of International Trade and Industry (MITI). The research park is owned by KSP Inc. (19%), Tobishima Corp. (30%), Nippon Life Insurance Co., Meiji Mutual Life Insurance Co., and Nippon Landic Co. (about 17% each). The facility serves four primary functions: (1) to lease space to established companies to set up their research facilities, (2) to provide laboratories facilities, (2) to provide laboratories and equipment for new or other companies to conduct research or to test concepts and ideas (Incubation Business), (3) to provide a very comprehensive materials characterization laboratory, and (4) to conduct its own basic research. In addition, KSP has a hotel and a number of technical and office support services. Laboratories of the established companies are housed in the larger of the two buildings, a 12-story structure. NEC, Fuji Xerox, and Fujitsu are a few of the companies with laboratories here. The Materials Characterization Laboratory is part of the Kanagawa High-Technology Foundation, which was formed by Kanagawa Prefecture to support and provide service to industry in the region. This laboratory is large and extremely well equipped. Our hosts said it was the best in Japan. The equipment in their surface characterization facility includes a Rutherford backscattering spectrometer, an X-ray photoelectron spectrometer, scanning Auger spectrometer, Fourier transform infrared spectrophotometer, and two electron probe microanalyzers. In addition, there is equipment for instrumental elemental analysis, thermal analysis, environmental testing, material structure analysis, and mechanical property testing. Their in-house research laboratories are part of the Kanagawa Academy of Science and Technology, which was established in 1989. At present, there are three research laboratories; three more will be established next fiscal year. They plan ultimately to have 10 laboratories. Each laboratory has a staff of seven to eight. In addition, they accept about 10 researchers from industry to be trained in research and a smaller number of graduate and undergraduate students. They presently have four foreign researchers from Korea and Taiwan, and next year one of the laboratories will have an American researcher. I visited the Molecular Spectroscopy Laboratory and the Biology Laboratory. The former is headed by Dr. Hiroo Hamaguchi. Some of the current work is on time-resolved Raman spectroscopic study of photochemical reactions and the development of an ultrasensitive multichannel Raman system. The Biology Laboratory is headed by Dr. Toshihiro Akaike; in his absence, Dr. Chia-Wun Chang briefed us on the group's work on molecular recognition membranes. They are working on a liver cell mimic and are designing membranes to be used as molecular sensors. Both laboratories are spacious and extremely well equipped, but because they are only a year old, there were no major accomplishments to report. However, the researchers are all young and enthusiastic, and I believe they will accomplish much in the near future. The third laboratory is headed by Dr. Norimasa Iida and is conducting research in support of the development of a two-cycle ceramic methanol engine. In commemoration of the first anniversary of the opening of KSP, an international forum was held on 24 October 1990. The forum was titled "Japan and the World in the 21st Century: Road to Creating a New Scientific and Technological Civilization and Japan's Role." The invited speakers were Gavriil Popov, Mayor of Moscow; Prof. Rustum Roy of Pennsylvania State University; and Keiichiro Hirata, chairman of the Japan Comprehensive Development Council (Sogo Kenkyu Kaihatsu Kai). Prof. Emeritus Shigeto Tsuru of Ishibashi University chaired a panel discussion with the speakers. In conclusion, KSP is truly an impressive and very well equipped research and development facility. There is a good mix of laboratories, conference rooms, meeting halls, and support services to encourage good interaction among researchers from various companies and organizations. Because the KSP laboratories are so new, there isn't much in-house progress to report. Also, I did not visit any of the laboratories of the established companies, so I cannot comment on how they interact with each other. What KSP does tell us is that the Japanese have ample resources and are willing to invest a great deal of that into encouraging R&D and developing new high-technology industries.--Sachio Yamamoto, ONRASIA DEQSOL AND ELLPACK: PROBLEM-SOLVING A number of integrated problem-solving environments for the solution of INTRODUCTION The numerical solution of partial differential equations (PDEs) was one of the earliest applications of electronic digital computers, and it remains the source of many challenging computational problems today. Such problems can be found in every scientific discipline. They range from computing the currents and fields in very large scale integration (VLSI) devices to determining the flow about an airfoil. by Ronald F. Boisvert and David K. Kahaner Due to the importance of mathematical models based on PDEs, and their general resistance to analytical treatment, the numerical solution of PDEs has been the focus of a great deal of research over the years. In spite of great algorithmic advances that have been achieved, very few general-purpose software packages for solving PDEs have appeared, and those that have tend to concentrate on very narrow problem areas. General-purpose PDE software is just extraordinarily difficult to build. This is due to difficulties in designing both solution algorithms and Since PDE problems exhibit such a The potential payoff for the development of such systems is great. Solving PDE problems by writing Fortran code is a very time-consuming and error prone task that often duplicates work of many previous programmers. Scientists and engineers may not be aware of the most appropriate algorithms to use, and hence the programs produced by these efforts may well be very slow or even produce erroneous results. Such difficulties are increased in the face of vector and parallel processing computers, which are very difficult to make efficient use of by the casual user. Using software parts from existing program libraries can alleviate some of these problems, but this still requires much low-level programming effort, even to solve easy problems. A number of researchers have made efforts to produce integrated problemsolving environments for the solution of PDEs (see the short bibliography at the end of this article). Two major systems that remain active foci of research are DEQSOL, a Japanese project, and |