How big is a mainframe?

Question

If you read about the history of computing, you'll hear how the first computers were "huge". You will often come across assertions that in the early days of commercial computing, a single computer would be "so big that it filled an entire building".

Now, poking around Wikipedia, I can find plenty of photos of old computers the size of an entire server rack, or several server racks. But I can't seem to find any pictures of a computer filling an entire room, much less a whole multi-story building.

Are these claims of a computer "filling an entire building" actually accurate, or is that a wild exaggeration? I can well imagine if you just paid a few million USD for a computer, you probably put it in its own special room with locked doors. But do any of these systems really fill a whole building? Do any of them really "fill" a whole room? Most pictures seem to just show a mostly empty room with cabinets across one wall.

Answer 1

In the 1980's a certain bank with its headquarters in Edinburgh has a problem with (IBM) disc storage that had to be kept online for live customer account information for branch and ATM machine operation that it ran out of city buildings to put the disc drives in.

Yes: Not just a building, but buildings. Luckily, just after that time radical developments were made in disc storage densities and the need for more real estate diminished, but computer floor space was a big issue at the time.

Here, also, is a picture of the machine room at Manchester Computer Science, containing on single machine, the MU5. This is just the processor, the peripherals and disc storage are in another adjacent room:

^{_{Source: http://www.cs.manchester.ac.uk/about-us/history/mu5/}}

It was rather large, but the lower floor computer room that contained an ICL 1906A was even bigger; and then there was the CDC 7600 and the Cyber 106 too.

An earlier machine was Atlas. Here is a picture of the large room containing only the processor of the London University Atlas; several other large halls contained the peripherals and storage:

_{^{Source: http://www.chilton-computing.org.uk/acl/technology/atlas/p010.htm}}

They were all very big power hungry beasts that took some real estate.

There are plenty of examples on the internet in the computer history archives.

Answer 2

But I can't seem to find any pictures of a computer filling an entire room, much less a whole multi-story building.

(Image taken form Centre for Computing History)

Well, for example look at this picture of a 4341 setup. This is a small entry-level mainframe of ~1980. With a believable setup for such small machine with:

• CPU (in the middle against the 'wall')
• Networking,
• 4 Tapes with Controller,
• 4 Fixed Disks
• 6 Removable Disks,
• Card Reader
• Card Punch
• One High Speed Printer
• One Small Printer
• One Operator Console
• One Terminal

It's a reasonable setup for this machine, and I'd certainly call that a room full.

(Image taken from Science and Technology Facilities Council)

A large 3090 of the same time will, as seen here, fill about the same space with just the CPU. And it'll definitely need more peripherals than just a handful of disks and tapes.

Are these claims of a computer "filling an entire building" actually accurate, or is that a wild exaggeration?

As usual it depends on the size of building you look at. A fully configured mainframe of the '70s or '80s, with adequate peripherals, can easy fill 1000 m² (~11,000 sq ft). Then again, companies using those kind of commercial computers usually had more than one machine.

Let's take a nice example of a mid to upper size bank system like I had a job with in 1981. They had a building the size of a Tesco Superstore (or one of these large DIY stores) to house 6 computers with all I/O and offices for machine operators and IT management. No user or any other department was located there. About ⅔ of that building was the machine room. 5 of them were used for daily business, while the 6th was a developer system. One of these five had the single job of operating a high speed optical reader, an awesome device ... anyway.

You see, they could get pretty big. A CPU (that's the mainframe term for the computer itself, processor, memory, memory interface, I/O controller and I/O interface - so without any peripheral device, not even a boot disk) occupied four to six 21" full-size racks (or more, for high-end systems), depending on the memory installed. These 1980s machines were already considered small, as the previous generation could have up to 10 racks just for the CPU. Later, around 1990, everything fitted in just 1-2 racks.

It was mentioned by AnoN that this (the 4331 setup) doesn't look like much and one could just as well stack everything against the wall. Maybe, if these were standard 19" mini computer components. But they aren't. They are not made up from small 1...4 HE components, to be operated from one side.

They are mainFRAMEs - Each rack (like) unit had two frames installed on each side, which could be swung out like a door (after opening the doors), Depending on the machine and usage, this frame was a single component - or in case of CPUs maybe only a part thereof. For more compact items, like memory, usually a half-frame featured one compact component. For maintenance both sides had to be accessible, that's why you'll (almost) never see a mainframe pushed against a wall, but with at least 95 cm walkway between units (*1). The closest to compare may be two 19" racks pushed back to back.

Similarly, peripherals also needed access from both sides - in addition to front access for operators, someone needs to change disks and tapes, refill paper or take the printouts. Disks at that time (*2) were top loaders. So no stacking either.

While these large metal cabinets may look somewhat like today's 19" racks, they are not. It's a completely different design. Much like Mainframes are.

Mainframes are in many ways more like an industrial production line, than today's computers. Today's machines have their origin in minicomputers which are more like a multi-tool for a handyman - extremely versatile but slow. Useful for everything, but excelling at nothing.

---

*1 - I started out in Mainframe service, and we really appreciated any additional space above the bare minimum.

*2 - near the very end of removable disks IBM introduced a line of full-height racks with 3 disk units in each to be pulled out and loaded.

Answer 3

Here's a picture of the "Strela" (arrow) computer (1954) Vacuum tubes, 2000 op/sec (on 43-bit fixed point), 150kWt, 300 m³

Answer 4

The excellent history book "AN/FSQ-7: the computer that shaped the Cold War" describes the size of a 1950's era radar monitoring air defence computer.

A single installation (a single computer) was built in a specially designed four-storey "blockhouse" - 74 feet high and providing 90,000 square feet of floorspace. A separate building housed the generators and cooling towers. The "computer" itself took 2 whole floors. One floor housed telephone equipment and the top floor was all the operator consoles.

Some stats: 10 technicians just to look after the computer. More than 100 display consoles. Three-shift staffing meant it needed around 627 people.

To quote from the book:

A typical installation required about 1 megawatt of electrical power to run the central computer with its about 55,000 vacuum tubes, 175,000 diodes and eventually even 13,000 transistors (all of which were spread over more than 7,000 pluggable units), weighing in excess of 275 tons, and occupying about 21,000 square feet of floor space.

Edit: if you want to be really picky, each AN/FSQ-7 installation was actually two individual computers, duplexed for fault-tolerance. But it was still a monster.

Answer 5

Here's the computer room of the Shuttle Mission Simulator (SMS), building 5 of NASA's Johnson Space Center, in the early 1980s. Most of the boxes are parts of a Sperry-Univac 1100 mainframe, but some are "intelligent controller" satellite computers, Perkin-Elmer 8/32s. There are also some Singer-Link Flight Simulation proprietary visual system cabinets and a unique interface device for the IBM flight computers.

There were two simulators in the building, so there are two of everything.

I was privileged to be a simulation software engineer on this beast from the early 80s to the late 90s. The Univac mainframes lasted until, I think, the early 2000s, when they were replaced with Silicon Graphics machines.

Answer 6

The main computer hall of the company I worked for in the 1970s and 1980s was about half the size of a soccer pitch - about 200 feet by 150 feet. That contained three IBM S/370 mainframes at one end, and the rest of the room was packed full of disk drives, stacked up to 6 or 7 feet high with narrow walkways between, with the outside walls lined with tape drives.

The power supplies and cooling systems filled the whole of the ground level, and the computer hall was the next floor up, built on a false floor to accommodate the wiring and plumbing for the water cooling.

To be fair, that was only half the complete building - one floor of the other half was an open plan area filled with punched card operators, and the other floor was occupied by programming teams.

The magnetic tape library occupied about half the machine hall area, on the third floor - basically, wall-to-wall racks of 12-inch tape reels, and a staff of tape librarians to make sure things didn't get lost!

At a later time there was also a Cray supercomputer in the main hall - though unless you knew where it was, it was almost hidden from view by all the rest of the kit.

Answer 7

Such rooms were not FULL in the sense of not being able to get any more stuff in them, like a storage locker might be, because you had to be able to get to the various pieces. There was room to walk between the various items, and open panels for servicing, although sometimes not very much. But they were full in the sense that almost all available, useable space was taken up by the machinery. You could not put in any more without blocking access to stuff you needed.

Answer 8

If you want to know exactly how big the classic 1960s and 1970s IBM mainframes were, see the System/360 Physical Planning and System/370 Physical Planning guides. They give complete measurements of every part of the system.

Consider the IBM System/360 Model 85, a large (but not the largest) mainframe from 1964. The processor itself took up 12 refrigerator-sized cabinets, weighing a total of 5.5 tons and occupying an area 28 feet by 24 feet.

The entire computer installation was much larger than the CPU itself, since you'd have card readers, disk storage units, numerous tape drives, I/O controllers, stand-alone cabinets of memory, and so forth. Adding 2 megabytes of memory to the system used the IBM 2385 Processor Storage unit, a set of 11 cabinets that weighed 7 tons and took up 23 feet by 23 feet.

To answer the original question, mainframes could fill anything from a large office to an entire floor. Filling an entire building was pretty unusual except for the SAGE air defense system or maybe an NSA installation.

Answer 9

One of the largest computers ever built was the SAGE system, built to gather information about surprise attack on the US. It filled a building. One might argue that SAGE was more than just one computer. But if you accept it as just one computer, it meets your criterion. Wikipedia Article

Edit: A closer look at the Wikipedia Article shows that SAGE was a collection of computers, and a network. Each computer was an AN/FSQ-7, which is mentioned in another answer. That computer was the largest ever built, taking up about 21,000 square feet of floor space. That's big.

Answer 10

Consider ENIAC. From wikipedia:

It weighed more than 30 short tons (27 t), was roughly 2.4 m × 0.9 m × 30 m (8 ft × 3 ft × 98 ft) in size, occupied 167 m2 (1,800 sq ft) and consumed 150 kW of electricity.

That's roughly building-sized.

Answer 11

A mainframe in the 1970's was not too different from today's desktop computers, except for physical size.

They had a CPU, which was about the size of a refrigerator.

They had disk drives, each of which were about the size of washing machines.

They had I/O devices, which often ran the size of a desk.

They had cooling, which was the size of a whole-house AC unit.

They had wiring to connect everything up. Just like the tangle under your desk except using higher voltage driving higher current.

Unlike your PC, they also had multiple tape drives each again the size of a refrigerator, and those tapes needed to be stored on racks somewhere.

So, if you scale today's CPU/memory/mobo to something the size of a refrigerator, you're looking at maybe 1000x. So, following that, visualize the space take up by 1000 PC cases, then double it to account for the tape drives, thicker wires (for higher currents and voltages) and extra AC airflow. Then double it again because people had to have access to all these components all the time.

Don't you think that would fill a very large room? Even if you're using a laptop as your basic measure, it's going to be bigger than a typical office cubicle.

Answer 12

I can't remember where I saw or put it, but sometime in the last couple of months I came across an installation for an early IBM system somewhere online, in scanned PDF form (maybe in the Bitsavers documentation archive?), most likely their first "RAMAC" hard-drive using rig plus all the associated access and computing hardware for using it at the heart of an electronic accounting system.

It's some pretty heavy engineering. The first parts of it lay out the minimum allowable, and recommended dimensions for the room that the drive, the computer, the operator console, their various power supplies and glue logic racks will be installed into, in plan form, mapping out their positions (a multi-cabinet U-shape with the free-standing operator console nestled within) plus the floor-level cabling runs and the space needed for operators to get around the back of the machines to access service panels etc that would otherwise be sandwiched against the walls. I can't recall the minimum dimensions, but the recommended space wasn't far off the entire floor area of my apartment (which is about the same size as several previous workplaces, encompassing different offices and even medical examination rooms), pegging it around a square 25ft / 7-and-a-bit metres on each side. The access space around the "back" of the cabinets was no more than 4ft even so.

And that's for a relatively minimal system. Bear in mind that the term "mainframe" comes from an old term for the central processor of a computer - each of the system's primary components (CPU, memory, I/O handler, power regulator, etc) would be built into one or more of those large "frames", of which the processor (made up of a great many individual, interconnected rackmount cards stuffed full of discrete components or, much like HAL 9000, small-scale integrated circuit packages) was of course the "main" example. The singular encompassing the plural, it came to imply any complete computer system built along similar lines, regardless of size...

And of course as well as the plans there were photos and an artist's-impression birds-eye view of the machine in use. The two or three humans in the picture didn't look ever so big compared to the computer, but it was still sort of human-scale, instead of being a monstrous factory-filling item.

Of course, that's not literally filling the room, but it is taking up enough of it, in awkward enough positions, that you can't really make practical use of the space that's left unless you install the machine in the corner of a considerably larger room and employed the alternative L-shaped layout instead.

A "minicomputer" setup, by comparison, would have been one small enough to only occupy one or two such frames, or within a wider but lower piece of furniture analogous to a mainframe's operator console, with the processor, memory, offline storage devices, power supplies, and maybe even the user interface (where it wasn't a separate compact terminal or a teletype) all fitting within that limit.

Incidentally, you make reference to mainframes being "like" rackmount servers... actually the comparison is more direct than you might first imagine. The format of 19-inch communications and server (etc) rackmount frames is pretty much exactly the same as that of, particularly, IBM mainframes and/or DEC minicomputers. I can't remember if it's a direct heritage or just convergent evolution, but in either case it's the same environmental limitations that moulded them - to whit, they have to be able to fit through a standard doorway without getting stuck or damage being caused to equipment or building fabric, even if that may mean temporarily taking the door off its hinges. And it also needs to get around tight corners in corridors, stairwells, etc whilst in transit to the target room and being turned to fit through the door. But at the same time you want to have the maximum amount of useful space within each frame to minimise the floor area taken up by it, to maximise the amount of tech you can cram inside each box, and minimise the number of boxes and thus the building and shipping costs. So the format quite rapidly converges on the largest rectangular box that will fit through doorways and can be moved easily through buildings made to a common human scale. I think at one point it was even commonly called IBM Frame Size. But in general it's about 24 inches wide, a shade under 7ft high, and about 30 inches deep...

(I've even seen distinctly non-mainframey equipment clearly built to similar guidelines - like the back-projection mobile interactive whiteboards made by Smartboard back before short-throw lenses were good enough to make slimline front-projection models practical. They were quite a bit wider across the beam of course, but measured just less than a doorway's width front to back - a little less than a mainframe in fact, so you didn't have to take the doors off - were just short enough to fit under the lintel when lowered to their minimum height, and the castors and V-shaped integrated trolley compensated for the width by making it somewhat easier to "lever" around difficult corners and wiggle into rooms from narrow corridors... I expect all-up size was something similar to the largest part of the RAMAC setup, ie the massive, ~36-inch, horizontal-spindle hard disc unit proper, which needed a crane to lift on and off of trucks and planes and sometimes demanded remodelling of buildings to get it installed... like, holes would have to be made in walls and then bricked back up again afterwards, floors preemptively reinforced, etc. With the console being almost as bad, but at least being lower-rise and a sensible weight. Though the Smartboard was way lighter and moveable by one person, whereas even the pure logic parts of a mainframe usually needed a team of two or more...)

Answer 13

If you read about the history of computing, you'll hear how the first computers were "huge".

This depends on what you call a "computer" and what you call "huge".

The Zuse Z3 (see Wikipedia article) is seen as the first binary computer; it was built 5 years before the ENIAC and looking at the photo I think it was not larger than a large wardrobe.

So it definitely did not need a whole building.

However before the Z3 there were non-digital machines:

The "Analytical Engine" (see Wikipedia article) was designed in the 1830s but never built. That computer would have worked with mechanic parts only (such as gears); the device would have filled a whole building while the memory (the equivalent of "RAM" in modern computers) would have needed most of the space.

I heard about a mechanical computer using ternary instead of binary arithmetic that has actually been built in the 1890s. Unfortunately, I don't know how large this device was.

... commercial computing ...

Commercial computers came later. According to Wikipedia, the first commercial computer (UINVAC I) needed about 36 square meters and weighted about 8 tons.

36 square meters is much less than a whole building.