Control Data corporation
The Control Data corporation (CDC) was the computer company which was famous as a pioneer of the super computer. Seymour Cray belonged through the 1960s and continued making the fastest computer. However, clay left the company in the 1970s and founded Cray Research (CRI) and took the world fastest title. CDC was one of nine major computer manufacturers of the 1960s (as for the others IBM, Burroughs, DEC, NCR, GE, Honeywell, RCA, UNIVAC).
Table of contents
A background and the origin: I am until 1957 by World War II
During World War II, I let the development of the machine which decoded the code that U.S. Navy gathered engineers, and was made with the Japanese military and mechanical code making device of the both German forces hit it. A team collected in Washington D.C. created that kind of machines. The navy felt uneasy about a security leak when I broke up this team in postwar war expenditures reduction and has begun to look for a method not to break it up.
And a solution was found at last. I was settled by having the company called Chase Aircraft of Minnesota Saint Paul take over engineers. The company lost most of the contracts with the military with the end of the war and was in trouble. Therefore I spoke it without explaining what a team was doing just I of I was saved very much for the military if I employed merely them. I doubted it at first, but the officer of the naval higher position came to come at every meeting and understood an earnest thing. The company agreed that the team was based in the military glider factory after all.
Engineering Research Associates (ERA) was established and at a glance in this way engaged in unrelated various projects until the early 1950s. One of those is world's first commercial stored program computer, ERA 1103 of 36 bits. This machine was developed for the navy and was intended to use it in decryption center. It was considered to be a problem that the navy performed "possession" of ERA substantially in the Diet in the early 1950s. As a result, in ERA, relations with the navy were quitted (including the raise in fund), and the owner of ERA sold a company to Remington Landes Corporation in 1952.
Remington Landes Corporation maintained a team of ERA and continued developing the new product. It is magnetic drum memory system of ERA that the company was interested most. It merged with Sperry soon, and the Remington land was Sperry rand, and the ERA section was absorbed in the UNIVAC section of Sperry. Technical talent was bought, and the people who came at first from ERA participated in various projects. However, the UNIVAC II () project left to ERA by UNIVAC was late for a delay, and was concerned; all approximately let condition be out of order.
The company's custom of the big company resigns as 合 trap and engineers of ERA which I felt and establishes Control Data in Minneapolis in 1957. The CDC establishment member picked William Norris to a chief executive officer with one accord. In addition, Seymour Cray was chosen as a chief designer, but was not able to join CDC because work about 1103-based navy tactics information system (NTDS) continued until it was completed.
An early design: Great plan of the clay
Because CDC sold the subsystems such as magnetic drum memory systems mainly, I started business. When clay joined in the next year, I developed transistor-based 6 bits small size machine "Little Character" immediately. This was a test for the ideas of the large transistor base machine which Cray thought about.
Little Charactor gained success, and, in 1959, they released CDC 1604 () for the transistor which made 1103 48 bits. The first finished product of CDC 1604 was delivered to U.S. Navy in 1960. With the opinion which added address (501 Park Avenue) of the first location of CDC to "1103" which Cray still developed as for the number called "1604" before [1].
CDC 160A () that scales it to 12 bits, and fell is released in 1960. This is often considered to be the world's first minicomputer. 160A was a shape of the standard office desk and was the design which was rare for those days. In addition, the new version CDC 3000 () series of 1604 architecture will be sold until the mid-1960s.
The clay started the design of the machine of the world's best performance. The aim is performance 50 times as large as 1604. To that end, it was required was bold, changing it, and a project took time (it takes four years really). The administration began to mind the trend of the project, and Cray offered own research institute establishment in 1962 because an eye of the monitoring became severe and said it when I resigned if not admitted. Norris consented to this, and the clay took a team. In the administration of CDC which included Norris, a promise not to be able to visit the research institute of the clay is exchanged unless I am invited [2].
Peripheral device business
Through the 1960s, I came to feel that Norris had to develop something serious to be opposed to IBM. Therefore I started a bold strategy that he purchased the peripheral device company and prepared a lineup. They set a 10% price cheaply from IBM and tried that I provided a 10% high-performance thing. This was not always easy.
One of the first peripheral devices has a magnetic tape unit. I wrote the peripheral device section to let the section share cost in this development other than an office, and a slight discussion occurred. The peripheral device section is without a method to profit if it supplies an apparatus to other sections with the price of cost price equivalency. The peripheral device section was settled instead by having you return a part of the profit when a peripheral device sold, and the afterward method settled.
A card reader, card punch, a tape drive (strike reamer), the thing of the all company design including the drum-type printer were developed afterwards. As for the printer business, a Chinese holly carburetor in the suburbs of Detroit produced facts at first, but establishes a joint venture later. Furthermore, the Chinese holly sold a possession stock to CDC and was admitted into CDC as a printer section.
Norris explored the input means for the punched card which IBM influenced. He decided to purchase Rainbow Engineering Corporation which was the pioneer of the optical character recognition (OCR) system. The idea was a thing to dispense with a punched card by reading the thing which I devoted myself to with the typewriter of the font that an operator was decided on commonly in OCR. A large quantity of information is included farther than one piece of punched card in one piece of typed page (the punched card has only one ぶんの information of the type basically). Therefore, it is saved the paper. However, this was the switch that was difficult than I thought. CDC performed an important role by an early OCR system, but OCR has not become the main data entry means to this day. The factory of Rainbow performed a shutdown in 1976, and CDC stopped the business, too.
When it was revealed that a punched card did not disappear immediately without progressing as an OCR project thought, it was approached by the need that CDC coped with this immediately. The punched card apparatus continued being made, but they had high cost price. Therefore I will purchase Bridge Engineering company having the apparatus which is cheaper, and is high-speed. The magnetic tape unit came to prepare it in the factory, too. It was made a spin-off, and the factory and the factory of the printer became joint venture Computer Peripherals Inc with NCR (CPI) later. This reduces cost by commonizing development; had a meaning. Furthermore, this increases ICL later, too. Furthermore, in 1982, a printer factory was sold to Centronics from there.
Norris presented the office which provided calculation service for the medium and small-sized business which could not purchase a computer in many places. This business did not produce much profit. I suggested to Norris that some managerial classes closed the office which was not released of the profit in about 1965. However, I ordered general economy instead without accepting the suggestion because Norris liked this idea.
It is the birth of the super computer CDC 6600
On the other hand, Seymour Cray and 34 engineers continued a design at a new research institute. In 1964, the result was released as CDC 6600 and was proud of performance more than 10 times in comparison with every machine which there was in the market. 6600 CPU's were comprised of the functional unit which was plural async and ten I/O processor was connected and took charge of the general task load of most. Therefore the CPU leaves common work such as a Hollerith card and the disk input and output to a controller and can devote myself to data handling. I recorded performance of about 1MFLOPS with the latest compiler when I wrote a code with an assembler performance of 0.5MFLOPS. It is the number that is wonderful when this thinks about the times. Version CDC 6400 and 2 processor version 6500 that cut down on performance were released.
FORTRAN compiler called MNF (Minnesota Fortran) is developed for 6600 in the University of Minnesota [3].
When 6600 was shipped, IBM paid attention to this new company. When the Thomas J Watson youth said, "I did what defeated our company where the small company where there are only 34 people holds several thousand even if I let this gate-keeper how", I am informed it. It is said that Cray which heard this said that it is, "there is an answer in the question". In 1965, IBM started a development project of high-speed machine ACS-1 () at 6600. 200 people were gathered for this project. This project produced the interesting architecture and techniques, but it was incompatible with System/360 of IBM. A turn was planned to become the System/360 compatibility, but therefore performance lowered, and the project was canceled in 1969 without being able to bring about any product. Therefore many engineers concerned with this resigned from IBM, and a brain was lost by the high-performance computer section of IBM.
On the other hand, IBM jumped ahead and announced new version Model 92 of the System/360 series of the performance equal to 6600. This machine did not exist, but did not cancel it until people waited for the release, and 6600 sales stopped (marketing technique named FUD today). Norris did not leave this lie and accused IBM of violation of antitrust law one year later. CDC obtained assets of IBM of the 80 million dollars equivalency after all [4]. Norris paid attention to Service Bureau Corporation (SBC) in that. The company owned a computer and offered computing power to other companies. This fitted in into the calculation service business that Norris pushed forward well [5].
In the process of 6600 designs, CDC carried out an SPIN project to provide a high-speed magnetic disk unit system in the system. It was unknown whether a magnetic disk replaced magnetic drum memory, and it was unknown which of a fixed model and the removable model was good for the disk in those days. Therefore I tested every choice by the SPIN project and finally brought about 28 inches of fixed disks type and pack type of 14 inches. Hard disk business started from an SPIN project becomes the main force of CDC afterwards.
CDC 7600 and 8600
In a month same as what won against IBM, CDC announced new machine CDC 7600. I assume clock frequency 6600 approximately 4 times (10MHz ⇒ 36MHz) and realize throughput more than 4 times as the whole system.
It is the adoption of the instruction pipe line and does the decisive factor of the performance enhancement when it can process at the same time plural orders because the all parts in the CPU process an order in sequence each. Therefore, as a whole, throughput improves, and the execute time of the program is shortened even if the processing time of 1 order is the same.
Because 1,969 years when 7600 came up were the economic recession period of the United States, the reaction in the market was not sweet. In addition, I produced a problem in reliability because it was a too complicated design. I included only the function that was incompatible, and most of the OS's were made newly and was basic that was complete with 6000 series. 7600 projects got the profit somehow, but it followed that I lost reputation of CDC.
The clay started CDC 8600. 8600 classified 7600 into four, a smaller housing basically. 8600 could work in higher clock frequency by reducing size, and shortening a signal path and should have improved performance in high-speed memory and the thing that I put together. However, 8600 took a manufacturing method from before. In other words, I soldered parts on a board. Because there were too many soldering points, the machine did not work even if one place became poor in contact and, as a result, never worked properly. If the fresh start of the design was necessary, I decided the clay.
STAR and Cyber
In addition to 8600 re-designs, CDC had a project called STAR. Gym Sohnton who worked with clay in 6600/6700 led STAR. The design of STAR processes only a mathematical operation by a pipeline having a long it very with a thing using the technique called a vector processor now. The mathematical operation was full of it and was speeded up, but was in this way later than 7600 other than it. Because it was rapidity of the mathematical operations that a customer of real CDC demanded, it was not considered to be the big problem. The vector of STAR had 65,536 length by 64 bits width.
1960s end game, CDC did not have the financial power as they let these two projects go at the same time. I talk with Cray about Norris, and the clay establishes Cray Research in 1972. Norris continued helping Cray afterwards. 8600 was canceled in 1974, and STAR was released as Cyber 203 in the same year. However, STAR was not able to give performance by the real user's program at all. Gym Sohnton leaves CDC and will found Network Systems Corporation.
CDC packages it on the basis of the 6600/7600 architecture again and sells it as the CDC Cyber () series. This was the main product of the 1970s of CDC. Hash-up version Cyber 205 of STAR was cast into the Cyber series, too, and this was high-performance unlike an original as such. However, Cray gave the machine which the technique of STAR such as Cray-1 took in at that time, and a computing performance was extremely high.
Cyber 200/205 which CDC released was performance to be able to fight against Cray Corporation on equal terms in the specific field. CDC participates in various projects of the customer and participates in project black project of the military. One of the result includes Cyberplus by CDC Advanced Flexible Processor (AFP) another name.
Directionality of another design includes "Cyber 80 "project. This can execute 6600 programs, but the architecture is totally new 64 bits. I aimed at the release of 1980. This expected the replacement by purchase of the user of 6000 series. The project delayed and was finally released in the different name.
In addition, I planned the switches from hardware dependence to a service-centered business form. For example, there is development of learning system PLATO using the computer. I worked on Cyber system and adopted the latest user interface including the terminal with the touch screen of the BitMap method.
I it to ETA systems, a hard disk and the distance of the oblivion
CDC was going to challenge performance again. However, I thought that Norris became stiffening too much as an organization to create the designs that CDC had good quickly. Therefore he established new spin-off company ETA systems in 1983. The design target of the ETA is performance 40 times as large as 10GFLOPS, Cray-1. The ETA was not able to achieve an aim, but I had laurels of the world's best performance in my hand temporarily, and there were the some sales in the next year, too. CDC thought about making up for a loss by selling ETA, but I discharged most employees in 1989, and the employee who remained employed it again in CDC.
On the other hand, a big company such as Japanese NEC and Fujitsu has begun to enter the market. The super computer market is not so big. CDC looked for other markets. It is a high-performance hard disk market to have found. At the mid-1980s, a hard disk began to be equipped with by a personal computer and was a promising market. CDC occupied the seat of the top by OEM from the 1970s through the early 1980s in the market of 14 inches of disk drives. I go into 8 inches of disk drive markets which シュガートアソシエイツ cut open quickly. I came to occupy a big position in a particularly high-performance hard disk. However, it is large-capacity, and a technological advance advances by the entry of the new company rapidly, and CDC comes to fall behind it. It is CDC to have developed ATA standard with Compaq and western digital.
Curiously CDC let a hard disk section make a spin-off in 1988 and had the company called Imprimis. The Seagate technology that demanded the clue to the high-end market in the next year purchases Imprimis. CDCのコンピュータ部門はコントロール・データ・システムズと改称、BTグループの傘下に入り、BTのグローバルサービス部門に吸収された。
CDCの中でもエネルギー管理部門は最も成功していおり、世界の発電所の25%で管理ソリューションを提供していた。1988年か1989年、これを Empros として独立させ、後にシーメンスに売却した。
CDCのサービス事業は1992年にスピンオフされ、Ceridian となった。同社は人事関係などのITアウトソース企業として成功している。Ceridian は1997年、CDC時代に政府関係のシステムを開発していた部門をジェネラル・ダイナミクスに売却した。
シティグループとの関係
シティグループは1998年、シティコープとトラベラーズ・グループが合併して誕生した。トラベラーズ・グループはサンディ・ワイルが1986年にCDCの子会社 Commercial Credit を買収し、それを中核として築いたものである。Comercial Credit は元々独立した企業だったが、1968年に敵対的買収をされそうになり、CDCが資金を提供して買収に対抗したという経緯があった[6]。
映画やテレビドラマでの登場
- 『地球爆破作戦』(1970): 冒頭に磁気テープ装置など初期のCDCの装置が出てくる。
- 『600万ドルの男』のパイロット版 (1973): OSIの研究所のコンピュータに "Control Data" というロゴがある。
- 『おかしなおかしな大泥棒』(1973): ライアン・オニールがCDCをクビになった主人公を演じている。
- 『トロン』(1982): フリンとローラがエンコム社に侵入するシーンに登場する計算機室で CDC 7600 が背景に映っている。これはローレンス・リバモア国立研究所で撮影された。
- 『ダイ・ハード』(1988): 計算機室に CDC Cyber 180 数台の実物と ETA-10 のモックアップが映っており、これらは全て CDC Demonstration Services/Benchmark Lab から提供されたものであった。撮影直前に別のメーカーにキャンセルされ、短期間で貸し出しが決まった。なお、マシンは返却後テストされ通常通り売られた。
- 『ゼイリブ』(1988): 主人公がサングラスを初めてかけたとき、CDC の広告を見ると、そこには "OBEY"（服従せよ）という文字が見えた。
脚注・出典
- ^ 当時のCDC従業員のインタビューである Reminiscences of computer architecture and computer design at Control Data Corporation の p.20 に記述があるが、それは単なる伝説だとして、p.21 に正式な由来の説明がある。
- ^ 招待されていないセールスマンがクレイの研究所を訪問した際の話が こちらにある。
- ^ Frisch, Michael (Dec 1972). "Remarks on Algorithms". Communications of the ACM 15 (12): 1074.
- ^ Oral history interview with Richard G. Lareau, Charles Babbage Institute, University of Minnesota.
- ^ "COMPUTERS: A Settlement for IBM", Time magazine, Monday, Jan. 29, 1973.
「訴訟を取り下げる見返りとして、CDCは多くのものを勝ち取った。約1600万ドルでIBMの Service Bureau Corp. を獲得。同社は所有するコンピュータのCPU時間を顧客に売ってデータ処理を行うサービスを提供している。ウォール街のアナリストは、同社の市場価値を約6000万ドルと見積もっている。さらにIBMは同社から5年間サービスを購入し、6年間は同様のサービスを立ち上げないことで、CDCが費やした弁護士費用1500万ドルを相殺するという。したがってIBMが支払う総額は少なくとも8000万ドルとなる。CDCのワンマン会長ウィリアム・ノリスは、大胆な訴訟が「我が社の歴史上最善の経営判断の1つになった」と語った」
- ^ Price, Robert (2005-11-11). The Eye for Innovation: Recognizing Possibilities and Managing the Creative Enterprise. 11 (1 ed.). New Haven, Ct: Yale University Press. pp. 168. ISBN 978-0300108774.
参考文献
- Lundstrom, David. A Few Good Men from Univac. Cambridge, Massachusetts: MIT Press, 1987. ISBN 0262121204.
- Murray, Charles, and John Wiley. The Supermen: The Story of Seymour Cray and the Technical Wizards behind the Supercomputer. New York: John Wiley, 1997. ISBN 0471048852.
- Price, Robert M. The Eye for Innovation: Recognizing Possibilities and Managing the Creative Enterprise. New Haven, Connecticut: Yale University Press, 2005. ISBN 030010877X.
- Worthy, James C. William C. Norris: Portrait of a Maverick. Ballinger Pub Co., May 1987. ISBN 978-0887300875
関連項目
外部リンク
- Control Data Corporation Records Charles Babbage Institute, University of Minnesota
- Oral history interview with William Norris Charles Babbage Institute, University of Minnesota.
- Oral history interview with Willis K. Drake Charles Babbage Institute, University of Minnesota.
- Oral history interview with Richard D. Conner Charles Babbage Institute, University of Minnesota.
- Oral history interview with Robert Price Charles Babbage Institute, University of Minnesota.
- Oral history interview with Eugene L. Baker Charles Babbage Institute, University of Minnesota.
- Oral history interview with Norbert R. Berg Charles Babbage Institute, University of Minnesota.
- Oral history interview with Mike Schumacher Charles Babbage Institute, University of Minnesota.
- Oral history interview with Charles F. Crichton Charles Babbage Institute, University of Minnesota.
- Information about the spin out of Commercial Credit from Control Data by Sandy Weill
- Private Collection of historical documents about CDC
- Control Data User Manuals Library @ Computing History
- Computing history describing the use of a range of CDC systems and equipment 1970-1985
- A German collection of CDC, Cray and other large computer systems, some of them in operation
This article is taken from the Japanese Wikipedia Control Data corporation
This article is distributed by cc-by-sa or GFDL license in accordance with the provisions of Wikipedia.
In addition, Tranpedia is simply not responsible for any show is only by translating the writings of foreign licenses that are compatible with CC-BY-SA license information.
0 개의 댓글:
댓글 쓰기