Sign in with
Sign up | Sign in
Your question

A Complete History Of Mainframe Computing

Last response: in Photo reports comments
June 26, 2009 6:50:26 AM

Mainframes arguably express man's highest achievement

But will it Blend?
June 26, 2009 7:39:22 AM

Wonderful article, thanks Tom's. =]

Killed a good hour of my day, and I very much enjoyed it.
Related resources
June 26, 2009 7:55:33 AM

Really cool. One observation, on page 7 I think the magnetic drum is rotating 12,500 revolutions per minute, not per second....If my harddrive could spin at 12,500 revolutions per second I'm sure it could do all sorts of amazing things like flying or running Crysis.
June 26, 2009 8:17:47 AM

Good article, however although not quite "Complete". There is no mention of Collosus (which was used to break Enigma codes from 1944) or The Manchester Small-Scale Experimental Machine (SSEM), nicknamed Baby, which was the world's first stored-program computer which ran its first program in June 1948.
June 26, 2009 9:11:22 AM

So the ABC was in fact the first mobile computer? The picture does show wheels under the table at least :)  But I guess netbooks are easier to handle, and have batteries
June 26, 2009 10:11:42 AM

I am with pugwash - its a good article but why does it seem like it is a bit US centric, no mention of Alan Turning or "Baby" and the Enigma code cracking machines of Bletchley Park
June 26, 2009 11:47:10 AM

Err what about the Zuse Z3?
June 26, 2009 12:48:16 PM

I agree with others, in that I am surprised that there was not even a mention of a Turing machine or other very early "computers".

Surely they qualified as Mainframes of their times?
June 26, 2009 1:11:58 PM

It's a shame that multiplication, addition and division benchmarks are not persistently noted throughout the article.

I know that now a days it's very much dependent on software design, but it would still be nice to follow the progression in terms of calculation power of the machines.
June 26, 2009 2:05:28 PM

25 pages??? i love ad block but damn this is annoying
June 26, 2009 2:20:03 PM

Where can we get an 80/80 of this article without all of the noise? No PDF?
June 26, 2009 2:27:28 PM

So.....can it play Crysis?

Out of curiosity, since its a metric I am more familiar with, what would the TeraFLOPS rating be in the newest and bestest from IBM. And how much would one of those bad boys set you back in the wallet.

Was a very educational and interesting article.
June 26, 2009 3:04:02 PM

"The 704 was quite fast, being able to perform 4,000 integer multiplications or divides per second. However, as mentioned, it was also capable of doing floating point arithmetic natively and could perform almost 12,000 floating-point additions or subtractions per second. More than this, the 704 added index registers, which not only dramatically sped up branches, but also reduced program development time (since this was handled in hardware now)."

Many of these statements are sure to be wrong. 1) For sure, it would not be faster at floating point than integer. 2) Index registers have to do with memory addressing, not branching.
June 26, 2009 3:33:07 PM

First, I agree with the title being misleading, and I apologize for it. It was never intended to be a complete guide, which would be virtually impossible. I don't know why that title was chosen.

The choice of computers was U.S. centric, because computers were U.S. centric. I chose only one mechanical computer, and it was made by IBM, since they were the dominant company. To add more computers would have been boring, and none of them were important technological milestones. So, while they might be specifically interesting to you, I was of the opinion too many computers from the same time frame would be boring. I almost chose the EDSAC over the EDVAC, but, went with the first design over the first implementation.

With regards to the index registers, "the IBM 704 added index registers and a “TSX” instruction that would branch to an address but leave the address of the TSX in an index register. A single unmodified branch could use that index register value to return."

Loops involve branching, branching involves memory addressing.

With regards to floating point vis-a-vis integer, you need to be more careful about what you're sure of. For one, multiplies and divides are generally slower, being much more complex. But, more to the point, this information is available directly from IBM.
June 26, 2009 4:08:52 PM

As one who live the mainframe era from the 2k machines for $500K...this story is incomplete without the story of the competition that was the force behind the commercial introduction at a furious pace of things we take for granted today.

Any mention of mainframes without the Honeywell H-800 series, the H200 series or Multics leaves out systems that have had a large influence on computing as we know it. The H-800 was one of the first multiprocessing systems of the late '50s, the H-200 was Honeywell's answer to the 1401 in the '60s and Multics merely contributed much of the hardware architecture for the Intel CPU used in today's PCs and foreshadowed UNIX and many of the development tools we use today. I saw no mention of GE and their 600-6000 series. And NCR. (Remember the term "BUNCH" as the competitors to IBM.)

So starting in the '50s, you should also have the history of the BUNCH woven in even to their demise. Not every great idea originated from IMB (though many did).

June 26, 2009 5:01:18 PM

What, and no mention of Lawrence Waterhouse and his work during WW2??? :p 
June 26, 2009 5:10:03 PM

Nice article, it was fun to review that history. I would have added mention of the groundbreaking Cray machines, especially the seminal Cray-1 (and it's successor X-MP) as the first "supercomputer." The X-MP looked like a futuristic chaise lounge with the main circuits in a center column surrounded by a circular padded bench. They were so arranged to reduce interconnecting wire lengths, as the speed was limited by the time it took electrons to travel through the interconnects...a speed of light limitation! The later Cray-2 was unique in that it was completely immersed in a bath of liquid Flourinert to cool the dense circuitry.
June 26, 2009 5:38:18 PM

Great Article! I learned something new today! I've never been so "into" the computer history before.
Thanks Rich Arzoomanian for writing this article.
June 26, 2009 5:54:19 PM

all jokes aside, this is the best tom's hardware article i have read to date. thanks for taking the time, effort and expense for putting it together.
June 26, 2009 6:36:38 PM

Very nice article. I learned very interesting things connected to development of computer architecture. My father used to work on some of these systems and he even has some of the original documentation stored somewhere. Also I have one intake fan from some IBM mainframe that my father removed ( I don't know the model) runs on 220V and uses 54 watts, still runs fine manly is used for cooling amplifiers or overclocked PC's:) 
June 26, 2009 9:29:33 PM

It's so funny to read, but our so called 'underpowered netbooks' probably where faster than any computer upto pic 22!
You read constantly of '20x faster', '30xfaster', but the Atom processor just beats the crap out of all those models save the last 2!
June 26, 2009 10:14:18 PM

I found that article very interresting and give it my "Must to read" mention. I've followed Tomshardware once 1997 and, i've also read the famous "Jumperless bios" on socket 7 article, but I found this article very instructive about mainframes. Mainframe are somewhat unknown by many pc user now, and I give you M. Arzoomanian, a five! Mainframe computers is the machine on the very top end of the computer ecosystem and the Atom processor have no point in common with the finest of one processor we could found in a mainframe. Remember that Atom cpu is a 5$ cpu, not a 5K and up cpu! To conclude, in Fallout 3, we can see some mainframe computers and use some terminals. And yes, I remember to have see and talk to a certain M. Von Neumann (lol!!!).

June 26, 2009 11:43:35 PM

June 27, 2009 12:24:38 AM

Nice article - the pictures are great. JV Atanasoff did his work at Iowa State, and the drum is the only remaining piece of the ABC computer left (they built a working replica a couple of years ago).

The Electrodata 215 should probably have been mentioned, along with the burroughs
b300 and its successors the 3500/4700/4800/4900/V-series line (a 1985 vintage v380 is still running at a city in california 23 years later, to be retired this summer).
June 27, 2009 2:56:46 AM

Wonderful article, Wish I could go back and witness all of these systems in action. Having old school mainframe work experience, would look great on a resume.
June 27, 2009 3:02:53 AM

"The choice of computers was U.S. centric, because computers were U.S. centric. I chose only one mechanical computer, and it was made by IBM, since they were the dominant company. To add more computers would have been boring, and none of them were important technological milestones. So, while they might be specifically interesting to you, I was of the opinion too many computers from the same time frame would be boring. I almost chose the EDSAC over the EDVAC, but, went with the first design over the first implementation."

The title should then read, "A Complete American History of the Mainframe."
June 27, 2009 3:34:31 AM

Thanks for the trip down memory lane.

While the PDP-8 was a notable machine, no one who worked on machines of that era would consider it a mainframe. The PDP-6 maybe, or the PDP-10, but not the PDP-8.

We use to joke that you could tell a minicomputer from a real computer by two criteria: (a) If it has a key-lock switch on the front panel it's a minicomputer; (b) If documentation doesn't come bundled at no extra charge, it's a minicomputer.

Do I detect an IBM bias? :)  True they set the trend, but it seems the CDC 6600 and B5000/5500 get short shrift; truly amazing machines in their own ways. Not to mention the first SDS/XDS systems. BTW the CDC 6xxx PPU's--aka CDC 160A's--were one of the inspirations for the PDP-8.
June 27, 2009 4:02:33 AM

wordswormThe title should then read, "A Complete American History of the Mainframe."

He already agreed, and said he didn't know why the title was chosen. Geez, *read*.
June 27, 2009 4:12:19 AM

Thanks for the article. I started in college on an LGP-30 with a 64k drum memory, a 10 cps paper tape reader, a teletype, no assembler or compiler, and an architecture much like the 360. Then jobs involving 1401, 1440, 1460, and several 1410s. Then 360 (and RCA Spectra 70/45s) and 370s and others before moving out of direct IT responsibilities in 1981. I've been into PCs since before then and ever since.

Just last Sunday I was trying to explain to my 8 year-old grandson what computers used to be like. (BTW, he and his 7 year-old brother "built" their own PC with me 6 months ago). How they filled a room, drop-ceilings and raised floors, cables like fire hoses, how you could literally see data stored on magnetic tape at 200 dpi, how many "refrigerators" a typical computer encompassed. And so on.

Now I have some pictures and tech specs to share.

Thanks for the memories to share.
June 27, 2009 4:13:30 AM

Awesome article, thank you!
June 27, 2009 4:46:08 AM

lamorpa -- Some 704 numbers...

Integer multiply/divide: 240us
Floating point add/subtract: 84us

Index registers as referred to at that time are not what you think of them today. From the IBM marketing literature: Three special electronic registers facilitate the writing of programs. In normal practice, many programs involve the repeated application of the same sequence of steps to data located in different parts of the memory.... These [index] registers increase the logical ability of the machine and at the same time provide for a reduction in the number of instructions needed to perform the proper shifting and manipulation of a given sequence.

The 704 (or the entire 70xx series) was often referred to as a "FORTRAN machine" because it was optimized for the nascent (at that time) FORTRAN language--or you could argue that FORTRAN was optimized for the 70xx. Whatever, it was optimized for scientific (i.e, FP) calculations.

To many today it might seem a bit bizarre, but at that time every gate counted, and you built machines optimized for a specific purpose. In this case, scientific calculations, which meant floating point, which meant building a fast dedicated FP unit. Dedicated as in "it doesn't share logic with anything else".
June 27, 2009 5:35:04 AM

Nice read. I just found out though that while the Mainframe market is still growing, HP just took some 250 customers away from Mainframes last year and lost 0 of their existing enterprise base the other way. This prompted a search into their premier offerings of HP-UX, OpenVMS, and NonStop. HP-UX has the same uptime objectives of 99.999% but has 4x the number of programs available. OpenVMS has an objective of 10-years with zero interruption and, like HP-UX, can run on multiple chip types for customizable performance. NonStop has triple lockstep differencing and wormhole routing (yeah... they actually call it that because it's mega-fast) which, when properly implemented, makes it the most reliable computing platform in existance... and very fast.

But the best part has to be that HP can concurrently (on their Superdome) run HP-UX, OpenVMS, Windows and Linux all in flexible virtual machines.
June 27, 2009 10:40:51 AM

Pity it didn't mention the Colossus at Bletchley park, which was a fully electronic computer running in 1943.

June 27, 2009 12:08:55 PM

Good article, but it's a shame you don't even mention Konrad Zuse's Z1 and Z3, which were fully automatic mechanical computers (including conditional branching and looping) at the beginning of the second world war, preceeding the Mark I by 4 years!
June 27, 2009 3:05:55 PM

Rich, congrats on one of Tom's best articles ever.

BTW, I love the personal tone of the article. Nothing like personal experience to add credability and expertise.

I also work at IBM, although through aquisition, I am amazed at the technology available in this company. Nothing compared to some of the technical talent I have seen, but very impressive nonetheless.

June 27, 2009 4:17:20 PM

Well, I for one am disappointed--while OpenSolaris is great, how about The Cell and gaming video on the mainframe?
June 27, 2009 10:55:29 PM

Nice to see an article that recognizes the ABC was the first electronic computer, not the ENIAC. Thinking the ENIAC is the first (electronic) computer is a common mistake. Though, I might be slightly biased since I took a few classes in Atanasoff Hall at Iowa State University (the Comp Sci building) as part of my EE curriculum.
June 28, 2009 12:41:31 AM

One of the photographs in the article is the wrong one; in the section on the Burroughs B5500, the picture is actually of the BRLESC computer, a one-of-a-kind computer built for the Ballistic Research Laboratory.
June 28, 2009 12:42:34 AM

Just for fun, is anyone know the Robotron german company ? Robotron is the is known to be the IBM of the Europe and they have produced some mainframe like the ESER one and DEC VAX 11/780 clones. It worth a look: (in german...)
June 28, 2009 2:38:02 AM

Good article up until the past page. I cannot remain quiet when these types of articles deviate into the "mainframe should be the end-all be-all platforms of the future" and the newest offering is beyond anything offered by a PC. And in fact, I enjoy that the author is comparing IBM's E64 to a Core7i, which not even Intel's Server/Business platform. (Again, Why are we seeing comparisons between mainframes and PCs with opinions are which is “better”? Isn’t this a history of the mainframe article?)

If the author would also like to compare advances in computing and the most advanced computing devices to date in regards to PCs/non-mainframes and Mainframes, the it should be with a rival batch operating platform( ). It is neither an x86 instruction set or CISC based system.

Again, I found the article interesting and informative right up until the point that the old school mainframe mentality formed by years of enduring PC users jokes about antiquated and forgotten platforms surfaced and started the rant we have all seen and heard before.
June 28, 2009 9:53:49 AM

Wait a minute, you said computing in decimal was inefficient for the ENIAC, but for the eServer it's not?

And talking about the 3081, you suddenly kept on dropping acronyms in there. Luckily I have a RS/6000, so I at least knew what DASD meant.
June 28, 2009 11:36:32 AM

Great article!I love these history-themed articles, because what we see today is just a glimpse of computr history.Well done.
June 28, 2009 12:08:01 PM

Great article. But the nit-picking comments you are receiving are a little disappointing. If you included everything that has been suggested, your article would have to be a book.

I started on the IBM 1401 with 4K memory in 1966 using Autocoder, which was really just assembly language. None of these old systems were powerful enough to run compilers. And the brief mention of words being alpha or numeric is also interesting. The IBM 7074 words could contain 10 numeric or 5 alpha characters. Instructions could be changed by flipping the signs on words from positive to negative - very efficient for programming with limited memory, but a bear to debug later, especially if you hadn't written the program.

The timesharing you mentioned is really the virtual predecessor of personal computers - an effort to give users a personal slice of the mainframe. It was started by the BUNCH but probably used most widely by users of IBM TSO and VM/CMS, and the VAX 11/780. Now mostly an obsolete technology.

A great article - I enjoyed it immensely.
June 28, 2009 12:20:54 PM

One other quibble about the article, about the way it compared the IBM 3033 computer and/or its successors to today's PCs. It might be noted that the IBM 3033 was an improved version of the IBM 360/168, and that was an improved version of the System/360 Model 85. The Model 85 introduced one important feature still used in present-day computers, cache memory. But all three computers were microprogrammed, like the 386 was. The earliest Pentium chips not only had on-chip cache, they were also pipelined (like the Control Data 6600 or the IBM 360/91), and used an advanced algorithm for division, and had hardwired control. Architecturally, the IBM 360/195 would be the best parallel to this. The modern technique of having a decoupled microarchitecture is not the same thing as microprogramming.
June 28, 2009 3:03:19 PM

How about the LEO (Lyons Electronic Office). The first computer to be used to handle everyday office work, and the first to be sold commercially.
June 28, 2009 7:24:44 PM

This is so typically American...

If somebody else beat you to an invention, you just ignore it so you can keep your self-centered world view.

An author form any other country would at least have given honourable mention to developments in other countries.

The British colossus was a bigger achievement in my eyes, than the ABC. Moreover, ignoring the first real computer by todays definition (Turing completeness), the Zuse Z3, is hardly understandable, except in a borderline nationalistic context.

I fully acknowledge that the development of computers in the 20th century was mostly an American thing, but history has no room for patriotism.