SUPERCOMPUTING IN JAPAN

The Japanese supercomputer market and the impact of Japanese supercomp uters on the world market has changed substantially since the last TOP 500 list was published. There are three important developments. First, the Japanese users and manufacturers rely more heavily on vector proc essing for high performance. Second, the Japanese manufacturers are in tegrating vector processing into novel scalable parallel computer arch itectures. Third, the Japanese government and private companies are in vesting considerable money into these very high performance computer s ystems. A major government sponsored supercomputer procurement program was completed in 1993 with 11 major procurements and a volume of more than a quarter billion US dollars. Different types of supercomputers were procured, traditional shared memory vector multiprocessors, RISC based MPP systems, and also the newly introduced vector parallel syste ms. The move towards RISC based parallel computers is not as strong in Japan as in other parts of the world. Both Japanese users and manufac turers continue to rely on vector instructions and large SRAM memories for computing power. At the same time MPP systems are investigated ve ry thoroughly in manufacturer's and user's research labs. Japanese org anizations like powerful computers. Japan has now only about 16% (82) of the systems in the list as opposed to 106 last year. But these syst ems contribute 27% (709 Gflop/s) to the worldwide installed R(max) per formance. The difference is mainly caused by a large number of older g eneration systems (NEC SX-2 and SX-1, Hitachi S-820/60 smaller, Fujits u VP2200 or smaller, and Cray) falling off the list. In spite of the r educed number of systems, Japan has now more R(max) per inhabitant (5. 7 Kflop/s) than the US (5.6 Kflop/s) and far more than Europe (1.3 Kfl op/s). On the first page of the list, the top 20 computers, 12 are fro m Japanese and 8 are American manufacturers. For the first time, two J apanese manufacturers are deviating substantially from their American competitors in their basic supercomputer design. Scalable vector paral lel computers are being introduced by both Fujitsu and NEC, with the d esign goal of combining the high single processor performance of vecto r processors with the high scalability of CMOS technology.