CDC 6600 - Some unpublished insider Information.

 


Index:
1. CDC 6000 Series Instruction Set.
2. CDC 6000 Series Transistors
3. The most critical Problem with the 6600 Serial Nr.3
4. The most significant Problem with the 6400 at CERN
5. CDC at CERN  Photos.
6. CDC Braclets and Tie Clips
7. CDC - Bendix Story
8. 6600 Instalations for the US Goverment.
9. The CDC 8600 Story
10. CDC 6600 to the rescue ... 

11. Extract from retating to the CDC 7700 Computer:




The CDC 6000 Series Instruction Sets


Transistors in the 6600

The CDC6600 contained over 4‘000‘000 Transistors of the Type 2N709:

This article was found in the Internet:
CDC Cray awarded Fairchild Semiconductor a $500,000 development contract to build a new transistor for the CDC 6600. Jean Hoerni met the specification by combining "gold-doping" - the addition of gold impurities - together with the new epitaxial deposition process. (1960 Milestone) The 2N709 (FT-1310) n-p-n device was introduced in July 1961 as the first silicon transistor to exceed germanium speed specification by combining "gold-doping" - the addition of gold impurities - together with the new epitaxial deposition process. (1960 Milestone) The 2N709 (FT-1310) n-p-n device was introduced in July 1961 as the first silicon transistor to exceed germanium speed.

_________________________________________________________________________________________________________

The most critical problem when building the CDC6600 Serial Number. 3

The specification the transistors required a switching between 4 and 5 nanoseconds
Now, there were two deliveries. The transistors in the first delivery had a switching time of 4.8 nanoseconds, the second delivery had a switching   of 4.2 nanoseconds. So, both deliveries were within the specifications.
Now, the problem was the following:
When an electric impulse passes the 20 transistors from the first delivery than the impulse arrives 1.6 nanosecond later than 20 impulse passed through a series of the 20 transistors from the second delivery.
When to two impulse after passing the 20 transistors are used to make a decision for the creation of the following, then the result is wrong

It took Seymour and his crew moth to find this out.


The most difficult technical problem with the CDC 6400 at CERN.

The machine ‘crashed’ very often. That means the screens of the operating console were suddenly white. The blank screen effect! Or other phenomena occurred, which were catastrophic to the overall functioning of the system.

 

Analyzing the ‘post-mortem-dumps’ showed that bits were missing in the part of a PPU memory in the area where the permanent program parts wass stored.

Easy to repair: Replace the core memories.
Did not help!
Replace the modules that receive the data from the memories.
Did not help!  Replace all the modules responsible for the PPUs one after the other. So. all 400 in chassis number one.
Did not help!
Replace all power supply elements in chassis one.
Did not help!

One night, shortly after midnight, the engineer in charge sat bored, tired, and humiliated at the control console. Like hundreds of hours before. He run the maintenance program that tests the memory of the PPUs. Up to now, this program never showed the slightest trace of an error. Restless and already falling into fatal thoughtlessness, he slowly rolled back and forth, monotonously sitting on the comfortable armchair equipped with wheels in front of the console.

And then it happened!!!

A huge spark hissed between the arm of the chair and the aluminum table of the console.

And the test program whistles!

A ‘fatal error’!

And it was the same type of failure that had been desperately searched for several months: Bits were missing in the permanent part of the memory in a PPU.
Now, immediately he tried to provoke the same error. To his wonderful relief, the same problem appeared again.

The greatest salvation!

Because with computers, their programs, and in computer science in general, a problem is solvable as soon as it can be reproduced.
Then it’s ‘only’ a matter of time before it is solved. How long this time then really is, is another question.

The cause of the lost bits was explained:
It was the static electricity in the room. Later measurements by a specialist from CERN showed values up to a voltage of 1000 volts!

As a first treatment, a fence, made of hemp ropes, was drawn around the entire computer. Visitors were no longer allowed in the computer room. The very few female programmers were instructed to wear only cotton skirts and no silky underwear.

As a further measure, a specialist from Minneapolis was flown in. The problem would not have occurred if all components in the machine had been correctly connected to the earth. The specialized engineer came. First of all, he wanted to see the reproduction of the problem himself.
He saw it!

Then he said: “I have to try that in the laboratory”. But there they had only wooden stools without wheels at the console.

 

 


CDC AT CERN

Photos 1973


CDC CEs against Bendix CEs (A non-technical Story)

After being hired, the CDC employees got a symbolic present. The Ladies a bracelet, the men a tie clip. These articles contained a colored (green?) stone. Every five years after employment, the employee received another bracelet or tie clip with a stone of a different color. This was very appreciated by the CDC 'folks', and they were proud of it, especially those with many year of CDC. When CDC bought Bendix, it also took over their Engineers. Those received the stones in the color of their years with Bendix. This fact upset the CDC People because most of the engineers coming from Bendix had worked for the company for many years.
PS: Bendix was founded in 1924,  CDC in 1957.
 


CDC:The bracelet and tie clip stones.

The hierarchy of the stones was red, blue and then white. Red was for 5 years with CDC, blue for 10 years, and white for 20 years with CDC.


6600 Installation in the US. (From the CDC in US federal Gov 1963 pdf


The CDC 8600 Story. (from Wikipedia)

In 1971, Control Data was undergoing a "belt tightening" due to the cost of an ongoing lawsuit against IBM, and asked all divisions to reduce their payroll by 10%. Cray begged to Control Data to exempt his division so he could get the 8600 shipping. When Control Data refused this request, he cut his own pay to minimum wage to solve the problem.

By 1972, it appeared that even Cray's legendary module design abilities were failing him in the case of the 8600. Reliability was so poor that it appeared impossible to get a whole machine working. This was not the first time this had happened: on the 6600 project Cray had to start over from scratch, and the 7600 was in production for some time before it started working reliably. In this case Cray decided the current design was a dead-end, and told William Norris (CDC's CEO) that the only way forward was to redesign the machine from scratch. The finances of the company were dangerous, and Norris decided that he could not take the risk; Cray would have to continue with the current design.

In 1972, Cray decided that he could not work under such conditions, and left CDC to form Cray Research (it was an amicable departure; Norris and other CDC staffers purchased some of the Cray Computer initial stock offering, which turned out to be a lucrative investment for them). For his new work he abandoned the multiprocessor concept, concerned that software of the era would be unable to take full advantage of the CPUs. He may have come to this conclusion after the ILLIAC IV finally entered operation at about the same time, and proved to have disappointing performance.

Team members convinced Norris that the 8600 could be completed even without Cray, and work continued at the Chippewa Lab. By 1974, the machine still didn't work correctly. Jim Thornton's competing STAR design had reached production quality at this point, and the 8600 project was then cancelled. In service STAR proved to have poor real-world performance, and when the Cray-1 entered the market in 1976, CDC was quickly pushed from the supercomputer market. An effort was made to re-enter the market in the 1980s with the ETA-10, but this ended poorly.


 

CDC 6600 to the rescue ...

By 1963, in an effort to avoid nuclear conflicts in the future, President Kennedy proposed the Limited Nuclear Test Ban Treaty, which was signed by the United States, the Soviet Union and Great Britain. But the U.S. soon bypassed the treaty’s restrictions using Seymour Cray’s CDC 6600 supercomputer. The computer was powerful enough to simulate thermonuclear 11 explosions without actually detonating a device. Thus the U.S. could continue its testing covertly, abrogating the treaty. If there was any direct example of a supercomputer affecting international affairs it was this. The United States could continue testing by the use of the premiere supercomputer on the planet, leaving the other signatories of the Test Ban at a disadvantage.

Extract from retating to the CDC 7700 Computer:

HEARING BEFORE THE SUBCOMMITTEE ON MANPOWER AND PERSONNEL OF THE COMMITTEE ON ARMED SERVICES UNITED STATES SENATE NINETY-SIXTH CONGRESS GRES FIRST SESSION 57-606

NOVEMBER 19, 1979

Senator BYRD. What importance do computers play in developing air defenses and antiballistic missile defenses?

ANSWER. Computer capabilities are key items in developing air defenses and antiballistic missile defenses. Computers are used to control radar beam pointing and energy management, radar signal processing to eliminate undesired signal returns and make estimates of target position and direction, correlate data from multiple radars for display in control centers, and to control interception firing parameters.

In air defense, the last two functions are key to defense effectiveness. A control center receives data from many radars and must first resolve the problem of one target being seen by several radars. Next, the control center computes and monitors the target's flight path. Finally, the control center commits interceptor forces to the target and positions the interceptor for identification or attack. The control center utilizes large, high speed computers to provide these calculations and display of information. The capacity and speed of the available computers limit the number of radar sites and targets that the control center can manage.

In antiballistic missile ( ABM) defense, large high speed computers are vital for radar control and signal processing as well as direct control of the interceptor missile. ABM defense can be divided into two parts : ( 1 ) detection and assess- ment of the size of the attack and the areas under attack, i.e. , the nature of the attack ; and ( 2 ) presentation of attack data to the ABM components of the defense for subsequent radar acquisition and interceptor missile flight computa- tion and control.

Detection of the attack is performed by large radars whose beam and energy management are often controlled by a large high speed computer. Upon initial detection of an object, the computer controls the radar beam to confirm detection and calculate the flight path of the object with sufficient accuracy to determine its target and to hand over the target to the interceptor system for subsequent engagement. The latter calculation requires tracking of each object

for several minutes while scanning for new objects. The capacity and speed of the available computers limit the number of objects that can be tracked at one time and thus determine the capacity of the detection and assessment system.

The detection and assessment system must also provide sufficiently accurate data for the ABM radar to acquire the target, once this " handover" occurs, data from the ABM radar is processed in a computer to calculate data to launch and control an interceptor missile throughout its flight. The computer controls both the radar and the interceptor missile flight, and thus is a major factor in the ABM system's probability of kill.

The performance of a nation's defensive systems is limited by the capacity and speed of their computers. Technological transfer of either state-of-the-art computers or the technology to build them provides another nation with the ability to increase the capability of their defensive systems, both in improved capability to destroy an attacker and in the increased size of an attack that may be successfully countered.

Senator BYRD. What potential value would the Control Data Corporation 7700 be if the computer were to diverted to Russia ?

ANSWER. The CDC computer's capability is well beyond that of the latest generation Soviet computers. It has the potential for very significant military

applications, particularly in the strategic area. Some of these applications are :

  1. Surveillance and Warning. -Application of a computer such as the CD 7700 would significantly improve the overall Soviet capability in this area.

The improved data handling capability would permit a more reliable and efficient tracking of each target, as well as increase the number of targets that

could be tracked simultaneously. The overall resolution of the surveillance and warning system would be upgraded through the increased data handling capability.

  1. Strategic Command and Control. -An effective automated command and control system is believed to be a major objective of the Soviet Union. This re- quires the processing of large amounts of data, and the communication of results in near-real time. The CDC 7700 would be extremely valuable for such an application. It could increase the amount of data that could be processed and significantly improve the near- real time capability of the Soviet command and control system.
  2. Antiballistic Missile Defense. -The CDC 7700 could significantly improve the performance of vital components of a Soviet ABM system. A large high speed computer such as the CDC 7700 could improve multiple radar control and signal data processing, as well as provide for better direct control of the in- terceptor missile.
  3. Nuclear Weapon Design .-High speed computers play an important role in the development of simulation models, test data processing and analyses of design alternatives-all critical to the development of an optimum nuclear weapon design. The CDC 7700 could be applied to these functions and would improve the Soviet R&D ability in this important area.

Senator BYRD. Do you have any evidence of such diversion ?

Answer. DIA has no evidence that this has occurred, however, the Commerce Department ( Compliance Division ) is currently investigating this possibility.

Senator BYRD. Has this computer been supplied to China ?

Answer. China has unsuccessfully attempted to legally import the CDC Cyber 76 computer. No evidence of illegal acquisition of the 7700 is available to DIA.

Senator BYRD. What are potential military applications of the CDC 7700 ?

Answer. Some of the potential military applications are surveillance and warning, antiballistic missile defense, command and control ( strategic and tactical ) , nuclear weapon design, satellite ground control and data processin stations and antisubmarine warfare signal processing. The response provided for the question concerning the potential value of the CDC 7700 diversion is also applicable to this question.


Kommentar hinzufügen

Kommentare

Jim Clark
Vor 3 Monate

Hans - Great info and info that I had never realized. Thanks so much. Jim

Bewertung: 5 Sterne
1 Stimme