file eugene-miya.html - docenting about the high end computers
To some extent, our museum is a collection of museums that were closed because they could not justify the floor space that they took up. The floor space was too valuable. So we have the former
You are very fortunate to be in the only location in the world that has almost every computer that Seymour Cray has designed - minus about 2 or 3 machines.
Cray 1 - Serial 6 Circa about 1975 - $8 machine
This is a one megaword machine - it is not byte addressable. There were 64 bits per word, so it would be about an 8 megabye memory if you took the number of bits. There were about 6 different operating machines which ran on this. Its freon cooled. It has a 12.5 nanosecond cycle time (80 megahertz).
[A] The first three applications of this machine were:
After that it branched out into other areas. For example, there was one here at AMES for a while to help design the airflow of wings, design cars, design the Macintosh, portable laptop cases, and many other things that you can kind of immagine.
[A] Basicly this was used by the government like this one here at Livermore. This one was actually at Livermore (about 45 miles away from here).
This machine was never located in American universities. There were three competing machines that were located at like Colorado State and Perdue - made by the company that Cray Research came out of called Control Data and it turned out that in the early 1980s people decided that there was kind of a crisis and people started to put these machines in universities - and the put out about 10 different machines of various kinds - and now there are only two of them left out there.
[A] You will see a console device on the other side of the room which is similar, but you connected a fast mini-computer and you moved data in and out using that kind of thing. Now depending on where you were, there were two schools of thought:
This was supposed to be 256 processors - in fact there were only 64 (one quadrant). What you see mostly there is was is called the Control Unit (if you will, the brain). And behind this CRT and a bunch of buttons is a Processor Unit on the ground, it says ILLIAC IV PEM. which is Processon Element. There were 64 of those and they were the parallel processing work horses. They had a small amount - a couple kilo-words - of memory and you had to fit all your problems on that.
We also had these large fixed head disks.
This was kind of a difficult machine to program
Where did my knowledge come from? Basicly I came from software. If I flub up with things having to do with hardware, that is why. I am not a hardware person per sey.
But I have been involved with many super computer kind of things.
There is nothing on this tape which can be construed as propriatary. I have signed non-disclosure agreements with a number of companies. You will hear none of that on the tape.
What happened was that in the mid 1980s I was asked by the government to do a survey of parallel processing so I have basicly since that point in time monitored the work going on in parallel processing and in supercomputing, and to a lesser extent what is going on in general in computer technology.
There are lots of people interested in this, not only NASA but the intelligence community and lots of other things of that type.