previous part
A Look at Worldwide High-Performance Computing and Its Economic Implications for the U.S.*
next chapter

A Look at Worldwide High-Performance Computing and Its Economic Implications for the U.S.[*]

Robert Borchers, Seymour Goodman, Michael Harrison, Alan McAdams, Emilio Millán, and Peter Wolcott

Robert R. Borchers is currently the Assistant to the Director for University Relations at Lawrence Livermore National Laboratory (LLNL). Before accepting this assignment in 1992, he was the Associate Director for Computation at LLNL. In that role, his responsibilities included overseeing all large-scale computing at LLNL. He has been active in the supercomputing community as a member of the Supercomputing Conference Steering Committee, the Program Advisory Committee for NSF Centers, and numerous other panels and committees. He is the immediate past and founding editor of Computers in Physics, published by the American Institute of Physics. Before coming to LLNL in 1979, Bob held professional and administrative positions at the University of Wisconsin, Madison, and the University of Colorado, Boulder .


414

Seymour E. Goodman is Professor of Management Information Systems and Policy and head of the Mosaic research group at the University of Arizona. He studies international developments in information technology and related public policy issues. Professor Goodman has chaired several national-advisory and study groups concerned with international computing, including the Committee to Study International Developments in Computer Science and Technology and the Computer Subpanel of the Panel on the Future Design and Implementation of U.S. National Security Export Controls, both under the National Research Council of the National Academy of Sciences. He is a contributing editor of International Perspectives of the Communications of the ACM and editor of technology and transnational issues for International Information Systems. Professor Goodman was an undergraduate at Columbia University and received his Ph.D. from Caltech.


Michael A. Harrison is Professor of Computer Science at the University of California at Berkeley. He received a B.S. and M.S. in electrical engineering from Case Institute of Technology in Cleveland in 1958 and 1959, respectively, and a Ph.D. from the University of Michigan in Ann Arbor in 1963. His activities on behalf of professional societies is extensive. He is currently a Director of the American Federation of Information Processing Societies and has served for four years on the Computer Science and Technology Board of the National Academy of Sciences. Professor Harrison is a consulting editor for Addison Wesley Publishing Co. and is an editor of Discrete Mathematics, Discrete Applied Mathematics, Information Processing Letters, Theoretical Computer Science, and the Journal of Computer and System Science. He has written five books and well over a hundred technical papers. Areas of research in which he specializes include switching theory, automata, formal language theory, protection in operating systems, electronic document systems, and programming environments. Currently, his work centers upon the creation of multimedia systems. Professor Harrison is the founder and Chairman of the Board of Gain Technology Inc.


415

Alan K. McAdams received his B.A. from Yale University and his M.B.A. and Ph.D. from Stanford University. He has taught at Cornell University throughout his academic career. Professor McAdams was a Senior Staff Economist with the President's Council of Economic Advisors from 1971 through 1972, with areas of responsibility in the economics of science and technology policy. From 1972 to 1978, he was a member of the NRC/NAS Advisory Panel for the Institute for Computer Sciences and Technology of the National Bureau of Standards. Professor McAdams chaired the Office of Technology Assessment Advisory Panel for the study U.S. Industrial Competitiveness: Steel, Electronics, and Automobiles (1981). He is a member of the American Economic Association and the Institute for Electrical and Electronics Engineers. His publications include "The Computer Industry," in Structure of American Industry, 6th edition (1982) , Economic Benefits and Public Support of a National Education and Research Network (1988); several monographs on electronic networks (1987–1988); and HDTV and the Information Age (1991) .


Emilio Millán graduated with a bachelor's degree in science and Russian from Dartmouth College, Hanover, New Hampshire. He is now a master's degree candidate at the Department of Computer Science at the University of Illinois-Urbana/Champaign, where he pursues his interests in machine translation and language analysis. Research programs that he currently pursues are being carried out in conjunction with Seymour Goodman and Peter Wolcott of the University of Arizona, Tucson.


Peter Wolcott received his B.A. (magna cum laude) in computer science and Russian from Dartmouth College in 1984. He is a Ph.D. candidate in the Management Information Systems Department at the University of Arizona. His specialties are the development of software and high-performance computing systems in the former Soviet Union and Eastern Europe. He is a member of the Institute of Electrical and Electronics Engineers Computer Society, the Association for Computing Machinery, and the Dobro Slovo Slavic-studies honor society .


416

Abstract

The Japanese are mounting a threat to the American position as preeminent producer of high technology. This threat has substantial implications not only for American high-technology industries and the high-performance computing industry in particular but for national security, as well. This paper examines the worldwide high-performance computing market in an attempt to place the U.S., Japan, and a number of other countries in a global context. The reasons for the erosion of American dominance are considered and remedies suggested.

A Brief Technical Overview of the Present-Day Landscape

The United States has historically been the dominant country in the world in terms of both supercomputer development and application. The U.S. has the lead in both vector and parallel processing, and Cray Research, Inc., continues to be the preeminent company in the high-performance system industry. Moreover, the wide spectrum of approaches employed by U.S. supercomputer developers has resulted in an extremely fertile research domain from which a number of commercially successful companies have emerged—CONVEX Computer Corporation, Thinking Machines Corporation, and nCUBE Corporation among them. The U.S. high-performance-system user base can claim a sophistication exceeding or roughly equal to that in any other country.

However, we are not alone. A number of countries have undertaken extensive research efforts in the high-performance computing arena, including the Soviet Union, Japan, and some in Western Europe. Others, such as Bulgaria, Israel, and China, have initiated research in this area, and many countries now employ supercomputers. In this section we examine some of the more substantial efforts worldwide.

The Soviet Union

The Soviets have a long history of high-performance computing. The USSR began research into computing shortly after World War II and produced functional digital computers in the early 1950s. The first efforts in parallel processing began in the early 1960s, and research in this area has continued steadily since then.

Soviet scientists have explored a wide spectrum of approaches in developing high-performance systems but with little depth in any one. Consequently, the Soviets have yet to make a discernible impact on the global corpus of supercomputing research. The Soviets to date have


417

neither put into serial production a computer of CRAY-1 performance or greater—only within the last few years have they prototyped a machine at that level—nor have they yet entered the worldwide supercomputer market. However, Soviet high-performance computing efforts conducted within the Academy of Sciences have exhibited higher levels of innovation than have their efforts to develop mainframes, minicomputers, and microcomputers.[*]

The BESM-6, a machine that is capable of a million instructions per second (MIPS) and was in serial production from 1965 to 1984, has been, until recently, the workhorse of the Soviet scientific community. The concept of a recursive-architecture machine with a recursive internal language, recursive memory structure, recursive interconnects, etc., was reported by Glushkov et al. (1974). The ES-2704, which only recently entered limited production, is a machine embodying these architectural and data-flow features. Computation is represented as a computational node in a graph. The graph expands as nodes are decomposed and contracts as results are combined into final results.

The ES-2701, developed at the Institute of Cybernetics in Kiev, like the ES-2704, incorporates distributed-memory flexible interconnects but is based on a different computational paradigm—there called a macropipeline computation—in which pipelining occurs at the algorithm level. Computation, under some problems, progresses as a wave across the processor field as data and intermediate results are passed from one processor to the next.

The ES-2703 is promoted as a programmable-architecture machine. The architecture is based on a set of so-called macroprocessors connected by a crossbar switch that may be tuned by the programmer. The "macro" designation denotes microcode or hardware implementation of complex mathematical instructions.

The El'brus project is the most heavily funded in the Soviet Union. The El'brus-1 and -2 were strongly influenced by the Burroughs 700-series architecture, with its large-grain parallelism, multiple processors sharing banks of common memory, and stack-based architecture for the individual processors. A distinguishing feature of this first El'brus machine stemmed from the designers' decision to use, in lieu of an assembly language, an Algol-like, high-level procedural language with underlying hardware support. This compelled the El'brus design team to


418

maintain software compatibility across the El'brus family at the level of a high-level language, which in turn enabled them to use very different architectures for some of their later models (e.g., the El'brus-3 and mini-El'brus, both very-long-instruction-word machines).

Most of the more successful machines, from the point of view of production, have been developed through close cooperation between the Academy of Sciences and industry organizations. One such machine, the PS-2000, was built by an organization in the eastern Ukraine—the Impul's Scientific Production Association. The PS-2000 could have up to 64 processors operating in a SIMD fashion, and its successor, the PS-2100, combines 10 groupings of the 64 processors, with the whole complex then being able to operate in a MIMD fashion. Although now out of production, 200 PS-2000s were produced in various configurations and now are actively used primarily in seismic and other energy-related applications. Series production of the PS-2100 began in 1990.

The development of high-performance computing in the Soviet Union is hindered by a number of problems. For one, the supply of components, both from indigenous suppliers and from the West, is inconsistent. Moreover, the state of mass storage is very weak. The 317-megabyte disks, which not long ago represented the Soviet state of the art, continue to be quite rare. Further, perestroika -related changes have caused sharp reductions in funding of several novel architecture projects, and a number have been terminated.

Western Europe

In Western Europe, while there has been no prominent commercial attempt to build vector processors, much attention has been paid to developing distributed processing and massively parallel, primarily Transputer-based, processors. Efforts in this realm have resulted in predominantly T-800 Transputer-based machines claiming processing rates of 1.5 million floating-point operations per second (MFLOPS) per processor, with up to 1000 processors and with RISC-based chips promising to play a sizable role in the future. To date, however, the Europeans have been low-volume producers, with few companies having shipped more than a handful of machines. Two such exceptions are the U.K.'s Meiko and Germany's Parsytec.

Meiko and Parsytec have proved to be the two most commercially successful European supercomputer manufacturers, with over 300 and 600 customers worldwide, respectively. Meiko produces two scalable, massively parallel dynamic-architecture machines—the Engineer's Computing Surface and the Embedded Real-Time Computing Surface—


419

with no inherent architectural limit on the number of processors. Among Meiko's clients are several branches of the U.S. military and the National Security Agency. Parsytec's two Transputer-based MIMD systems, the MultiCluster and SuperCluster, are available in configurations with maximums of 64 and 400 processors, respectively.

Lesser manufacturers of high-performance computing include Parsys, Active Memory Technology (AMT), and ESPRIT—the European Strategic Program for Research and Development in Information Technology.[*] The U.K.-based Parsys is the producer of the SuperNode 1000, another Transputer-based parallel processor, with 16 to 1024 processors in hierarchical, reconfigurable arrays. AMT's massively parallel DAP/CP8 510C (1024 processors) and 610C (4096 processors) boast processing speeds of 5000 MIPS (140 MFLOPS) and 20,000 MIPS (560 MFLOPS), respectively. Spearheaded by the Germans, ESPRIT's SUPRENUM project has produced the four-GFLOPS, MIMD SUPRENUM-1 and is continuing development of the more powerful SUPRENUM-2.

The Europeans have proved themselves as experts in utilizing vector processors as workhorses. Vector processors can be found in use in Germany, France, and England. Though the Europeans have been extensive users of U.S.-made machines, Japanese machines have recently started to penetrate the European market.

Japan

Japan is maturing in its use and production of high-performance systems. The Japanese have elevated vector processing to a fine art, both in the case of hardware and software, and are producing world-class systems that rival those of Cray. Moreover, the installed base of supercomputers in Japan has climbed to over 150, the number of Japanese researchers working in the realm of computational science and engineering is growing, and the quality of their work is improving.

The first vector processors to emerge from Japan, such as the Fujitsu VP-200, generated a lot of excitement. Initial benchmarks indicated that these early supercomputers, with lots of vector pipelines—characteristic of the Japanese machines—were very fast. The Fujitsu machine was followed by the Hitachi S-820 and then the Nippon Electric Company (NEC) SX-2, which was, at that time, the fastest single processor in the


420

world.[*] These machines also boasted many vector pipes, as well as automatic interactive vectorizing tools of high quality.

Recent Japanese announcements indicate that the trend toward greater vectorization will continue. The NEC SX-3, for example, employs a processor that can produce 16 floating-point results every three-nanosecond clock cycle, a performance that amounts to more than five GFLOPS per processor.

It merits mention, however, that while Japanese high-performance computers compete well in the "megaflop derby," their sustained performance on production workloads remains unknown. Huge memory bandwidth hides behind the caches of these Japanese machines, and the memories are a fairly long distance from the processors, which probably inhibits their short vector performance.

Parallel processing is not, however, being ignored in Japan. The Japanese have a number of production parallel processors now to which they are devoting much attention. In at least two areas of parallel processing, the Japanese have made significant progress. Most, if not all, Japanese semiconductor manufacturers are using massively parallel circuit simulators, and the NEC fingerprint identification machine, used in police departments worldwide, represents one of the largest-selling massively parallel processors in the world.

The Japanese recently have begun showing signs of accommodating U.S. markets. For one thing, Japanese manufacturers are exhibiting some willingness to accommodate the IEEE and Cray floating-point arithmetic formats, in addition to the IBM format their machines currently support. Secondly, some machines, notably the SX-3, now run UNIX. These and other existing signs indicate that the Japanese seek not only to accommodate the American market but to aggressively enter it.

The software products available on Japanese supercomputers and the monitoring tools available to scientific applications programmers from Japanese vendors appear to be as good as or better than those available from Cray Research. Consequently, applications software being developed in Japan may be better vectorized as a result of the better tools and vendor-supplied software. Further, Japanese supercomputer centers seem to be having little, if any, difficulty obtaining access to the best U.S.-developed applications software.

While the U.S. appears to be preeminent in all basic research areas of computational science and engineering, the Japanese are making


421

significant strides as the current generation of researchers matures in its use of supercomputers and a younger generation is trained in computational science and engineering. The environment in which Japanese researchers work is also improving, with supercomputer time and better software tools being made increasingly available. Networking within the Japanese supercomputing community, however, remains underdeveloped.

The American, Soviet, European, and Japanese machines and their parameters are compared in Table 1.

The Japanese Challenge and "McAdams's Laws"

We now shift focus and tone to take up a number of the economic and political issues associated with high-performance computing by employing "McAdams's Laws" to examine the nature and possible impact of the Japanese challenge in the high-technology market.

Introduction

At the end of World War II, the U.S. gross national product equaled over half of the gross product of the entire world. During the post-World War II period of American economic and military hegemony, the U.S. pursued a national policy that favored activities designed to contain "world communism" over the interests of its domestic economy. Starting from the position of overwhelming predominance, these choices seemed necessary and obvious.

Since that time, much has happened in the world to clarify our perceptions. World communism was not only successfully contained over the last 40 years, but today, in many nations, communism and socialism are being abandoned in favor of democracy and capitalism. The fall of communism in Eastern Europe and elsewhere is viewed by many as a harbinger of a "victory" over world communism and a demonstration of the superiority of American-style laissez faire capitalism to other economic systems. However, there are also many who believe that in its efforts to contain communism, the U.S. may have brought its economy—especially its high-technology sectors—to a position close to ruin.

Law 1—
That Which Is Currently Taking Place Is Not Impossible

The perception of U.S. dominance as assured and perpetual is severely flawed. The U.S. may soon cease to be the world's commercial leader in the field of supercomputers and has rapidly lost ground in other areas,


422
 

Table 1. Parameters of Various High-Performance Systems


Machine


Peak Performance

Number of
Processors

Year of First Productiona

American

     

CRAY X-MP

   0.87 GFLOPS

4

1983

CRAY X-MP

     2.7 GFLOPS

8

1988

CRAY-2

       2  GFLOPS

4

1984

CRAY-3b

      16 GFLOPS

16

N/A

CRAY C90

      16 GFLOPS

16

 

Soviet

     

BESM-6

        1 MIPS

1

1965 (1964)

ES-2701

    530 MIPS

48

N/A (1984)

ES-2703

        1 GIPS (32-bit)

64 macroprocessors

N/A (1985)

ES-2704

    100 MIPS

24 computational

1990 (1980)

   

48 communications

 
   

12 switching

 

El'brus-1

12–15 MIPS

10

N/A (1979)

El'brus-2

      94 MFLOPS

10

1985 (1984)

El'brus-3b

    6.4 GFLOPS

16

N/A (N/A)

El'brus-MKP

   560 MFLOPS

1

1991 (1988)

Electronika-SSBIS

   450 MFLOPS

2

1991? (1990)

PS-2000

  200 MIPS (24-bit)

64

1981 (1980)

PS-2100

   1.5 GIPS (32-bit)

640

1990 (1987)

European

     

Parsytec MultiCluster

 

64 (max.)

 

Parsytec SuperCluster

 

400 (max.)

 

AMT DAP/CP8 510C

  5,000 MIPS
    (140 MFLOPS)

1,024

 

AMT DAP/CP8 610C

 20,000 MIPS
     (560 MFLOPS)

4,096

 

ESPRIT SUPRENUM-1

         4 GFLOPS

   

Japanese

     

Fujitsu VP-200

        4 GFLOPS

1

1983

Hitachi S-810/20

  0.63 GFLOPS

1

1983

Hitachi S-820/80

      3 GFLOPS

1

1988

NEC SX-2

   1.3 GFLOPS

1

1985

NEC SX-3

    22 GFLOPS

4

1990

Fujitsu VP-400E

   1.7 GFLOPS

1

1987

Fujitsu VP-2600

      5 GFLOPS

1

1990

a   For Soviet machines, "year of first production" is not necessarily a good benchmark, so in parentheses appears the year that prototype testing and refinement began. N/A indicates that the machine never entered serial production.

b   Projected values.


423

as well, including machine tools, consumer electronics, semiconductor-manufacturing equipment, and high-performance semiconductors. Even areas of U.S. strength, such as aircraft and computer hardware and software, may soon be at risk unless strong action is taken. American competitiveness, much less dominance, in these and other high-technology areas can no longer be assumed.

Alarms have been sounded at many levels. The National Advisory Committee on Semiconductors, in its recently released second annual report, refers to the semiconductor industry as "an industry in crisis" and urges the federal government to act immediately or risk losing the semiconductor industry in its entirety and with it, the computer industry, as well. The Administration itself has just identified 22 critical technologies vital to U.S. military and economic security, a list of technologies virtually identical to those identified earlier and individually by the Departments of Defense and Commerce as vital to the future of the U.S. in world geopolitical and economic competition.

A concerted effort on the part of the Japanese, combined with complacency on the part of American industry and unfavorable trade conditions between the U.S. and Japan, have brought about this situation in which spheres of U.S. industry have lost former dominance and competitiveness in certain international markets. The world has changed. The U.S. is no longer predominant.

Japan:
Vertical Integration, Keiretsu, and Government Coordination

In 1952, when the Japanese became independent, they set a goal, embodied in the motto, "We will match the standard of living in the West." At that time, over half of the Japanese population was engaged in subsistence agriculture, and yet, by improving their output-to-input ratio, they were able to improve their productivity. Since then, the Japanese have moved from subsistence agriculture into light manufacturing, into heavy and chemical goods, and into the higher-technology areas. Today, as a result of an innovative corporate structure and governmental industrial orchestration, the Japanese have positioned themselves to become the dominant suppliers of information technologies to the world.

Japan today has a different, more sophisticated structure to its economy than our own. The major Japanese firms producing computers are all vertically integrated, meaning that a strong presence is maintained across the spectrum of computing machinery—from micros to supercomputers—and in all allied technologies: microelectronics, networking,


424

consumer electronics, etc. In contrast, there is only one vertically integrated company in the computer field in the U.S.—IBM.

Table 2 illustrates this situation. The three leading Japanese supercomputer firms—NEC, Fujitsu, and Hitachi—are integrated all the way from consumer electronics to microcomputers, minis, intermediates, mainframes, and supercomputers. All are large-scale producers of semiconductors. In contrast, U.S. firms producing semiconductors are either "merchant" suppliers, with the bulk of their sales and earnings coming from the sale of semiconductors on the open, merchant market, or "captive" suppliers, such as IBM and AT&T, which produce semiconductors only to satisfy internal demand. Further, when previously successful merchant suppliers have been purchased and merged into large U.S. companies to become both captive and merchant suppliers, they have uniformly gone out of the merchant business.[*] Usually they have shut down completely. No U.S. captive supplier has become a successful merchant supplier to the market. Japanese firms, however, do both successfully. They are captive suppliers to themselves, and they are merchant suppliers to the market. This suggests something amiss in our system or between our system and that of the Japanese.

 

Table 2. Market-Segment Participation by Selected U.S. and Japanese Manufacturers of Supercomputers and/or Semiconductors

 

Cray


DCD/ETA

IT, Intel, Motorola


AT&T


IBM

 

NEC


Fujitsu


Hitachi

Supercomputers

x

x

   

xa

 

x

x

x

Mainframe      Computers

 

x

   

x

 

x

x

x

Intermediate      Computers

       

x

 

x

x

x

Minicomputers

 

x

 

x

x

 

x

x

x

Microcomputers

       

x

 

x

x

 

Consumer      Electronics

           

x

x

x

Semiconductors      (Merchant)

   

x

     

x

x

x

Semiconductors      (Captive)

     

x

x

 

x

x

x

a IBM is reentering the supercomputer market.


425

Japanese firms are not only integrated across virtually every aspect of semiconductors and computers but are also prominent members of Japanese industrial conglomerates, or keiretsu . The major Japanese keiretsu groups are each structured around a major bank.[*] The Bank of Japan supports these banks with differential interest rates for loans to "target" industries.

From their keiretsu structure, Japanese corporations get diversification, risk reduction, and a lower cost of capital, which allows them to maintain a favorable output-to-input ratio. Additionally, while competition in their home market among the keiretsu forces reduced costs and improved quality, these firms will cooperate when operating in foreign markets or when facing foreign firms in their home market. Should they temporarily fail to cooperate, the Japanese government steps in and reestablishes "harmony," especially in their dealings with outsiders, or gaijin . Thus, the Japanese have not only a team but a team with a strategy.

The U.S.:
Rugged Individualism and Trade-War Losses

In contrast, American industry has rejected a close working relationship with the government and insists on that truly American concept of rugged individualism. Whereas this arrangement has at times resulted in extraordinarily rapid growth of the American high-technology industries, it has also resulted in an uncoordinated industrial environment in which poor decisions have been made.

Economic policy decisions must be made with respect to certain economic relationships, which can be easily illustrated formally. Four variables are required for this somewhat oversimplified example:

q = quantity of output;

p = price of products;

w = wages per hour; and

i = number of hours worked.

If five computers are sold (q = 5) for $1000 each (p = $1000), total revenue will be $5000 (qp = $5000). If the wage rate of the work force is $10 per hour (w = $10) and the input of work hours is 500 (i = 500), the cost to produce the machines is also $5000 (iw = $5000), and no profit is realized:


426

More generally,

 image

If we now divide both sides of this relationship by ip , we get

and canceling, we get

This ratio, in a nutshell, illustrates what is needed to break even. On the right-hand side of the equation are two factors expressed in dollars—w , the hourly wage and p , the price of the computers—whereas on the left is the ratio of q , machines produced, to i , input hours. The ratio of q / i is the rate of output per unit of input and represents what economists call the (average) production function of the process. It is a relationship determined by the technology. If the technology is the same in two countries, this ratio will be (roughly) the same for those countries. A country with lower hourly wages, w (e.g., $5), will be able to charge a lower price, p (e.g., $500) for its product and still break even. That is, with the same plant and equipment (or production function) on the left-hand side of the equation and low wage rates on the right-hand side, low prices will be possible for the products from the low-wage countries. If, however, wages are very high (say $20), then the ratio of w / p will require that p , the product price, also be high (in our example, $2000).

International competition is as simple as that. It is neither "good" nor "bad," it is simply inexorable. The simple relationship demonstrates why so many jobs are being lost by the U.S. to developing countries. Their low wages make it possible for them to produce products from relatively stable technologies more cheaply and thus charge lower prices than can we in the U.S.

There are solutions to this problem, but the U.S. has generally failed to implement them. For example, it may be possible to improve U.S. plants and equipment (and/or its management) so that employees are more productive and thereby achieve greater output (q ) per unit of input (i ). This would justify


427

a higher wage (w ) in relation to a given unit price (p ) for the output. This has been a major tenet of Japanese strategy for decades. Another solution involves producing a higher-quality product for which consumers will be willing to pay more, thus justifying the higher wages paid to the work force, since a higher p can justify a higher w .

Because the U.S. has as its long-term objective to maintain or increase its relative standard of living, then one or both of these strategies are required. Even then, a way must be found to inhibit the rate of diffusion to low-wage economies of an innovative, highly productive production process and/or of product quality innovations. Only in these ways can higher wages—and thus a reasonable standard of living—be sustained in our economy over the long haul. The implications of these facts are pretty clear; they are very much a part of our day-to-day experience.

A reasonable economic development strategy for the U.S. must be in the context of these major forces influencing outcomes worldwide. The major forces aren't definitive of final outcomes, but they do establish the limits within which policies, plans, and strategies can be successful.

The need to respond to the low prices offered by other countries has been recognized in this country for many years. The solutions attempted have largely been ineffectual quick fixes, in essence trying to catch up without catching up rather than facing up to the imperatives of improving product quality and productivity in general.

Law 2—
You Don't Catch up without Catching Up

Our response to the present challenge has, to date, included a miscellany of wishful thinking, "concession bargaining," and manipulating monetary factors. Concession bargaining sought to cut the wages of U.S. workers producing high-technology goods—and thus their standard of living—so that the U.S. could match the prices that low-wage countries are offering. When "concession wages" didn't work, the U.S. then decided to find another financial gimmick that would permit us to lower the price of our goods in world markets.

The U.S. decided to cut the exchange rate in half. At the bottom, the dollar was worth 120 yen, while before it had been worth 240 yen and more. In effect, this introduces a factor (in this case 1/2) between the world price before the change and the world price after, while leaving the U.S. domestic equation unchanged. Given that we have over a $5 trillion economy today, cutting our exchange rate with the world in half amounts to giving away $2.5 trillion in the relative value of our economy. The $30


428

billion improvement in the balance of trade due to the lower world price of our goods yielded a return on our investment of only a little over one cent on each dollar of value lost. This is not an intelligent way to run an economy.

Trade:
"Successful" Negotiations and "Potato Chips"

Another major difficulty for U.S. high-technology manufacturers is the asymmetry in U.S. and Japanese market accesses; U.S. markets are wide open through our ideology, whereas Japanese markets are not. The U.S. government, which historically has shunned "intervention," has been reluctant to "start a trade war" and insists that U.S. firms are better off standing alone. The government, whose policies might be stated as, "It doesn't matter if we export computer chips or potato chips,"[*] has been practicing what can only be called unilateral disarmament in international trade battles.

Law 3—
When Two Countries Are in a Trade War and One Does Not Realize It, That Country Is Unlikely to Win

The Japanese markets in target industries are, have been, and are likely to remain closed. In the U.S., we refer to such a phenomenon as "protectionism." This market was protected when the Japanese were behind. This market was protected while the Japanese caught up. This market remains protected even now after the Japanese have achieved substantial superiority in many of its products.

These facts violate conventional wisdom that equates protectionism with sloth. Clearly, that has not been the case for the Japanese. Protectionism for the Japanese can't be all bad; it is possible for protectionism to work to the benefit of a nation; it has for Japan. Today, the U.S. has a weekly trade deficit with Japan of about one billion dollars—almost half of which is spent on automobiles—which feeds into the Japanese R&D cycle rather than our own.

Protectionism has long been the name of the game for Japan, and the U.S. has an extremely poor track record in trade negotiations with the Japanese. There are general trade principles understood by both the U.S. and Japan, some of which are embodied in specific, written trade agreements. An important one is that each side should buy those technological


429

products that the other side can produce more economically. The Japanese, however, routinely violate these principles whenever it is convenient to do so, and the U.S. does not pressure the Japanese to meet their obligations. Seventeen negotiations between the United States and Japan in semiconductors have been almost completely ineffectual. These general comments are illustrated in Figure 1, which is also known to the industry as the "worm chart."

Figure 1 shows the share of the Japanese semiconductor market held by U.S. producers over the period 1973–1986 in relation to a series of "successful" negotiations between the United States and Japan to open the Japanese semiconductor market to competition by U.S. firms. It is startling to note that the U.S. market share has dropped by approximately one per cent—from 10 to nine per cent—in the wake of these "successful" negotiations.

During the early period, Japanese firms were no match for their American rivals. By the early 1980s, they had caught up in many areas and were already ahead in some. By the end of the period, Japanese firms had established general superiority. Yet, throughout the entire period, and irrespective of the relative quality of the U.S. versus Japanese products,

Figure 1.
The Worm Chart: U.S. share of Japanese semiconductor market.


430

the U.S. market share has remained virtually the same. It could be characterized as a "worm," lying across the chart at the 10 per cent level.

There's something about the way the Japanese manage their economy that has led to a constant U.S. share of its semiconductor business in the range of approximately 10 per cent. Given the multitude of variables involved, it is a virtual impossibility that markets alone could have brought about such a result.

Remedies

Law 4—
An Important Aspect of Change Is That Things Are Different Afterward

It is not difficult to see that no panaceas exist. Various remedies to ameliorate the situation, however, do.

The present-day situation must be recognized as a crisis. The U.S. government and, to a lesser degree, industry have failed to recognize this as a crisis. Fixing that which is wrong now is many times more difficult than would have been the case just a few years ago.

The Japanese market will have to be pried open for American products. When President Bush appointed Carla Hills to the post of U.S. Trade Representative, he gave her a crowbar to emphasize that very point. To do so, the Japanese must be held to their obligations under current trade agreements, and legislation intended to bring about a more equitable trade situation, perhaps along the lines of the High Performance Computing Initiative, should be passed. Should these measures fail, the U.S. might be wise to consider assuming some vestiges of protectionism, at least as the means to pry open the Japanese market.[*]

More qualified people are needed in Washington to attend to the problem. Both NSF and the Defense Advanced Research Projects Agency are trying to recruit such individuals, with both experiencing difficulty.

The industry has changed, and American industry must change with it. Many of the lessons to be learned in this case come directly from the Japanese. For example, to survive these days, a super computer company must be vertically integrated and must generate money to be invested in R&D.


431

More money must be invested in R&D. Another key to industrial survival today is a market share to generate money for investment in research and development. R&D spending in Japan went up 14 per cent during one recent 12-month period, alone, and four or five Japanese companies have R&D budgets that exceed the entire budget of NSF, including NEC, whose budget exceeds NSF's by one-third.

U.S. industry must improve its productivity, change its values, and technologically catch up. A top-down approach to the solution of eroding U.S. leadership in the area of high-performance computing will not work. Fooling with the exchange rate won't work, nor will any number of additional stopgap measures. Changing our values, above all else, means abandoning our short-term view in terms of industrial planning, education, consumer buying habits, and government involvement in industry. Incentives must be introduced if industry is to be expected to assume a long-term view. Unfortunately, industry in this country has a very short-term view. They won't take the technology that is available in the universities. The good technology that is developed there is being siphoned off to Japan, where there are interested people. Meanwhile, industry occupies itself worrying about satisfying the investors at the next shareholders' meeting.

Education—at all levels—must be improved in this country. The decline of public elementary and secondary education is well-documented and demands both increased governmental spending and fundamental changes in values. Similar reforms are necessary at the university level, as well. Improving education in America, however, can not be accomplished in short order. Sy Goodman:

The educational issue, long-term as it is, is still absolutely critical. The U.S. university community, in my opinion, is overly complacent about what it thinks it is, relative to the rest of the world.

Industry should investigate the possibility of government involvement, perhaps to the point of coordination. Summarily rejected by the U.S. government and industry, governmental coordination of Japanese industry by the Ministry of International Trade and Industry has been instrumental in Japan's postwar rise to its current position as a high-tech industrial superpower. Studies cited earlier show that all relevant elements of the public and private sectors are now agreed on those areas in which the U.S. must succeed if it is to remain a world-class competitor


432

nation. The U.S. must put aside adversarial relationships among government, industry, and workers.

The Future

When one looks to the East these days, Japan is not the only competitive nation on the landscape. The newly-industrializing countries (NICs) of East Asia (including Singapore, Hong Kong, Taiwan, and South Korea), while not known primarily for technological development, have exhibited proficiency in building components, peripherals, and systems of increasing complexity. It is not inconceivable that these countries will produce high-performance systems in the future. The NICs, however, unlike Japan, are not members of the Coordinating Committee of Export Controls. Worldwide availability of supercomputers from these countries could have a substantial impact on U.S. national security. If and when this time comes, the U.S. government should not be unprepared to address this matter.

In short, and in conclusion, both industry and government have a large stake in the continued health, if not dominance, of America's high-technology sectors, including the supercomputer industry. Both also have important roles to play to insure this continued health. Further, the industry and government need not work toward this goal in isolation from one another.

References and Bibliography

M. Borrus, "Chips of State," Issues in Science and Technology7 (1), 40-48 (1990).

D. H. Brandin and M. A. Harrison, The Technology War , John Wiley and Sons, New York (1987).

Federal Coordinating Committee on Science, Engineering, and Technology, "West European High Performance Computer Suppliers," FCCSET memorandum, Washington, DC (1991).

C. H. Ferguson, "Computers and the Coming of the U.S. Keiretsu," Harvard Business Review90 (4), 55-70 (1990).

V. M. Glushkov, V. A. Myasnikov, M. B. Ignat'yev, and V. Torgashev, "Recursive Machines and Computing Technology," in Information Processing 74: Proceedings of IFIP Congress 74 , J. L. Rosenfeld, Ed., North-Holland Press, Amsterdam, pp. 65-71 (1974).


433

M. A. Harrison, E. F. Hayes, J. D. Meindl, J. H. Morris, D. P. Siewiorek, and R. M. White, Advanced Computing in Japan , Japanese Technology Evaluation Center panel report, M. A. Harrison, Chm., Loyola College, Baltimore, Maryland (1990).

E. F. Hayes, "Advanced Scientific Computing in Japan," in Far East Scientific Information Bulletin NAVSO P-3580, Vol. 15, No. 3 , Office of Naval Research, Far East APO, San Francisco, pp. 109-117 (July-Sept. 1990).

S. Jarp, "A Review of Japan and Japanese High-End Computers," in Far East Scientific Information Bulletin NAVSO P-3580, Vol. 16, No. 2 , Office of Naval Research, Far East APO, San Francisco, pp. 59-79 (April-June 1991).

A. K. McAdams, T. Vietorisz, W. L. Dougan, and J. T. Lombardi, "Economic Benefits and Public Support of a National Education and Research Network," EDUCOM Bulletin23 (2-3), 63-71 (1988).

National Research Council, Global Trends in Computer Technology and Their Impact on Export Control , S. E. Goodman, Chm., National Academy Press, Washington, DC (1988).

C. V. Prestowitz, Jr., "Life after GATT: More Trade Is Better Than Free Trade," Technology Review 94 (3), 22-29 (1990).

C. V. Prestowitz, Jr., Trading Places: How We Allowed Japan to Take the Lead , Basic Books, New York (1988).

P. Wolcott and S. E. Goodman, "High-Speed Computers of the Soviet Union," Computer21 (9), 32-41 (1988).

P. Wolcott and S. E. Goodman, "Soviet High-Speed Computers: The New Generation," in Proceedings, Supercomputing '90 , IEEE Computer Society Press, Los Alamitos, California, pp. 930-939 (1990).

P. Wolcott, "Soviet and Eastern European Computer Performance: Results of Benchmark Tests," Mosaic technical report 1991-003-I, Tucson, Arizona (1991).


435

previous part
A Look at Worldwide High-Performance Computing and Its Economic Implications for the U.S.*
next chapter