IBM-7094

The IBM 701 stored 2 instructions per 36 bit word,
the rest stored 1 instruction per 36 bit word and were somewhat upward compatable (code written for a lower number machine might execute correctly on a higher number machine).

Hmmm - I think the Computer History Museum IBM-1620 Restoration Crew said that the 1620 air cooled memory box had a thermostat set of 104 degrees F. - and the unit would think about executing programs until the memory box was up to temperature?

I've been contacted about the potential availability of a complete IBM 7094 system in Rosedale, Australia. It's owner is Susan Vigors. Included are docs, tapes, punched cards, etc.. Also two 1401s.

Here's what Susan has written about the collection so far. Please contact her at < vigors at iinet dot net dot au > and please cc John Deane (Australian Computer Museum Society) and I (or whole list) if interested.

[The computer collection] started with David's purchase of the Weapons Research Establishment (WRE) IBMs at auction, following a suggestion by his cousin (then Purchasing Officer for WRE) that it would be prudent to keep the Defense computer in working order for a while in case a need for it arose. He added that its scrap value was unlikely to be recognised by the scrappies, and additionally it was the first computer on which David worked in the late 50's and early 60's. By the way, the 1401's we used at WRE as front end consoles for the 7094, which is why they were sold in the same lot. To this very large computer (I believe the largest ever in the southern hemisphere) was subsequently added a collection of computers over the years including calculators, manuals and accessories, with a view to one day assembling a computer museum on the property. We are in the process of converting the property to a tourist centre, but lack resources to initiate or manage the museum. So - any ideas? The shed where they are stored is needed for tools and equipment, but I have so far restrained my sons from sending them for scrap, partly on the grounds of their value far exceeding anything the scrap yard will offer.

The computers were initially stored securely in our barn, but the 7094 drum was so heavy it broke the floor boards. So after a few years they were banished to the machinery shed with a concrete floor, high iron roof and walls and iron sliding doors (like an aircraft hangar). They were kept dry, but not covered, so that dust, insects and rodents over the years have deposited on the tops of the cabinets, and infiltrated where doors were not securely shut. In the last 10 years my sons have pushed the cabinets close together and stacked smaller computer equipment on top of the 7094 to make space for a workshop in half the shed, so photos are hard to come by. There is no catalogue of the hardware and software. ... We have a set of manuals for the 7094 equipment, paper tapes and card decks for boot-up, tapes for the tape drives, and I'm not sure what software is included. ... We never planned to [restore the system], as the cost of power and air conditioning would have made it quite impractical. We envisaged a working tape unit, possibly driven by a 1401 or a more modern PC, and perhaps a tape or card reader to demonstrate the older technology.

Regards,

- Robert

• The IBM 7094 and CTSS
• "Computer Structures: Readings and Examples", (1971 edition) by C. Gordon Bell & Allen Newell, Part 6, Section 1, Chapter 41
• The IBM 7094 and CTSS by Tom Van Vleck

There is a 7090 "someplace" which may be going to a museum "someplace". To aid that planning and possible restoration, John Van Gardner, Feb 8, 2011, has scanned his copy of 7090 Power Supply Control and Distribution - 223-6904 4 megabytes John Van Gardner wrote "I never worked on or saw a 50-Hz 7090 but it raised some questions. The normal 7090 had its own 400 Hz mechanical generator mounted in the 7608 frame. It was driven by a very large 208 volt 3 phase 60 Hz motor. I seem to remember it was about a 50 hp motor. It was directly coupled axially to the generator. The ones I worked on were the brush types that had very large carbon brushes on the slip ring to excite the field winding. We replaced these brushes once a year and pumped grease into a Zerk fitting on each bearing until grease came out a hole at the bottom of the bearing housing. The 7608 frame did a great job of containing the noise of the motor and 400 Hz hum of the generator. You never realized this until you opened one of the covers and the noise hit you. The motor and generator created a lot of heat and there was an exhaust blower to pump it out the rear of the top cover. Some installations had the 7608 outside the computer room. The generator was manufactured by General Electric and any repairs to it had to be done by them. After they decided to get into the computer business they didn't seem to want to help us in a reasonable amount of time. We found out the best thing to do was get a loaner 7608 from IBM while GE fixed the customers generator. "All the DC power supplies in each frame ran off the 400 Hz but there was a long squirrel cage blower on the bottom of each gate that ran off 60 Hz. These blowers were probably the most troublesome parts in the 7090. There were multiple squirrel cages on the shaft and it was cog belt driven. They were noisy and the belts would strip or break. The bearings would fail. There were air flow sensing cards in the row above the blower to shut the machine down if a blower failed."

Some Q&A, Stan Paddock asking, John Van Gardner responding:

> How did the 7090 boot? Cards, tape, other? There is a Load Cards and a Load Tape button on the 7151 Console. The Load Tape button causes the necessary boot commands to be sent to tape drive #1 on channel 1. The Load Cards button causes the commands to be sent to the 711 Card Reader on Channel 1. > What language was used? Everything was in machine language which descended from the 704 then 709. There was an assembly program, Fortran which compiled a machine language object program and later on there was Cobol. The instruction set was very extensive and you could make the machine do anything you wanted to. I loved to write programs in machine language. There was a 729 tape drive diagnostic called 9T55 that would print out a graph depicting the performance of the prolay arms. > Was the 7090 code generated on the 7090 or on another machine like the 1401? The 7090 came before the 1401. The 7090 was compatible with the 709 and that was one reason IBM won the contract for the BMEWS computers. They were able to debug the programs on a 709 while the 7090 was being developed. An annotated picture of WRE in Australia

More Q&A, same questions from Stan Paddock, different view point from Lyle Bickley

> How did the 7090 boot? Cards, tape, other? It could boot from Cards, Tape and Disk (with a small loader). At Lockheed, Sunnyvale we used almost entirely tape.. > What language was used? Lockheed used mostly Fortran and Assembly with some COBOL. > Was the 7090 code generated on the 7090 or on another machine like the 1401? The 709x card reader was seldom used. We copied cards to tape on a 1401 and then mounted the tape as input on the 709x. Our "operating system" was "IBSYS" (really a "monitor" program). "Job" and other control cards would be used to create a batch jobstream which could consist of Fortran, Assembly, Loader, Sort, Edits, etc. There were typically many jobs loaded on to a single tape. Output was seldom to a 709x printer - rather the "printer" output was sent to tape. That tape would then be processed by a 1401 doing a tape->printer job. So the 1401 did the slow I/O, freeing up the 7090 to concentrate on computation. IBM also released a multiprocessor 7040/709x combination that was truly incredible. Because 709x channels were so expensive, it was actually cheaper to configure tape drives and disks on a 7040 with a fast processor to processor link between the 7040 and the 709x. In this configuration, the 7090 had no I/O channels attached - only the processor to processor link! This created an amazingly fast and cost effective system. (I developed a Fortran compiler written in Fortran on such a system in the development lab in Poughkeepsie.) ------ following comment by Ed Thelen ------------------ The above combo system was called HASP Fortunately (for their customers) superceeded in - some Bouroughs systems, - and General Electric 6xx systems (using GECOS) - and Control Data 6xxx systems (using SCOPE with the imbedded peripheral processors (PPs)) by "spooling" where the same processor did the fast processing, outputing blocks to system disk and - specialized sub-processes in the same computer read from the system disk and output to the slow devices such as printers and card punches. IBM finally was able to do the above in the 360 MVS operating system, I think. At GE and CDC, we felt really sorry for IBM customers.

Lyle Bickley responding:

The 7090/7040 feature was called "Direct Couple". See the paper at: http://www.softwarepreservation.org/projects/os/dc.html Here's the relationship to HASP: http://www.redbug.org/dba/sharerpt/share79/o441.html