The Story of the Z8000

Because the Seventies progressed and the primary 8-bit private computer systems began to turn into well-liked, one factor quickly turned clear. Moore’s legislation meant {that a} new era of extra highly effective 16-bit microprocessors would quickly be wanted.

Whoever received the battle of 16-bit architectures could be well-placed to dominate the microprocessor marketplace for this era and past.

In pole place was Zilog. The Zilog Z80 had been one of many main 8-bit microprocessors, soundly beating Intel’s designs. Its founder, Federico Faggin, had led the event of the primary business microprocessor after which early 8-bit microprocessors at Intel. Zilog was owned by and provided with plentiful capital by oil big Exxon, then one of many world’s most respected firms. Faggin was additionally early to recognise the approaching transition to the 16-bit microprocessor era.

But Zilog’s 16-bit design, the Zilog Z8000, is a footnote in expertise historical past. It was shortly and comprehensively crushed out there by each Intel’s 8086 and Motorola’s 68000. This was though Intel’s 8086 was, by many measures, a design that was markedly inferior to the Z8000.

How may this occur?

On this submit, we’re going to have a look at the story of the Z8000 and study why it failed. It’s a narrative that has classes for us even in the present day.

The early days of Zilog had been a triumph for Federico Faggin. He had constructed the primary business microprocessor, the 4004, while at Intel. He had then gone on to guide the event of the Intel 8080, the microprocessor that had helped begin the private pc growth.

However he fell out with Intel’s administration and left Intel in August 1974. Seeing the chance to create a devoted microprocessor firm, he based Zilog in 1975 with Ralph Ungermann and shortly gained vital funding from oil big Exxon, then trying to diversify into the fast-growing data expertise sector.

Zilog’s first product, the Z80, designed to appropriate with, but additionally considerably improve, the 8080, turned an enormous success. The Z80 discovered itself on the coronary heart of a brand new era of dwelling and enterprise private computer systems. The Tandy TRS-80 and Sinclair ZX Spectrum each used the Z80. Maybe extra importantly, the Z80 powered a variety of computer systems working Digital Analysis’s CP/M working system which dominated the enterprise computing market.

The story of the Z8000 begins early within the historical past of Zilog.

Even while improvement of the Z80 was underway, Federico Faggin noticed that his agency wanted to begin on improvement of its 16-bit successor. Masatoshi Shima who had labored with Faggin on the 4004, the 8080 after which had adopted him to Zilog was busy at work on the Z80. In any occasion, the successor could be more likely to be radically completely different to the Z80 and so Zilog wanted somebody with extra intensive expertise within the improvement of extra highly effective pc architectures.

Faggin recruited Bernard Peuto in March 1976 from Amdahl, the maker of IBM appropriate mainframe computer systems. Peuto was interested in the chance to create his personal structure from scratch, quite than merely construct designs based mostly on another person’s structure. Peuto had beforehand carried out intensive research of assorted points of structure design, with analysis papers printed on instruction size, code density and velocity.

Right away after becoming a member of Zilog, he began work on the specification of the Z8000.

Early on, some essential selections had been made which might set the course for the design of the Z8000 and would assist to find out its destiny.

Firstly there was the choice as to what number of pins (or electrical connections with the remainder of the pc) the brand new processor would have. This seemingly mundane resolution would have an important impression on the structure of the brand new processor. As Faggin would later say, it turned a case of the ’tail wagging the canine’.

After contemplating and rejecting, on price grounds, a package deal with 64 pins, the crew determined to go together with two choices. One cheaper model would match right into a 40-pin package deal. This model would have its reminiscence deal with sizes restricted to 16-bits and so would solely be capable to use 64 KB of reminiscence. A dearer 48-pin model would have entry to a 23-bit or 8 MB deal with area.

To assist this method, the structure could be based mostly on segmented addresses. Any reminiscence deal with could be made up of two elements – a 16-bit ‘inside phase’ deal with and an deal with for the beginning of the phase – which might be added collectively to generate the total deal with.

The 40 pin model, often called the Z8002, would solely function inside a single 64 KB phase. The 48 pin model often called the Z8001 would have entry to as much as 128 distinct segments of as much as 64KB.

There was a catch although. The Z8001 would want a further chip, often called a reminiscence administration chip (or MMU), to completely implement this segmented method. The Z8001’s MMU, often called the Z8010, would add extra price and complexity to the implementation of extra highly effective Z8000 based mostly programs.

Segmented reminiscence additionally made issues harder for some kinds of software program. Shifting from 8-bit programs was simple. In actual fact, Zilog constructed software program that may enable builders to mechanically translate their Z80 meeting language code right into a model that may run on the Z8000. Nevertheless, some rising lessons of computer systems and purposes, notably people who included Graphical Person Interfaces, would want entry to bigger steady blocks of reminiscence than was attainable with a design with 64 KB segments.

If creating the Z8000’s specification was tough, then it wasn’t helped by the connection between Peuto and Shima. It could be Shima who would implement the structure, and so it was essential that they labored intently collectively on the specification of the design. Based on Faggin:

“The 2 had been speculated to work intently collectively, however Shima was much more opinionated than Bernard and each time they met they appeared to conflict.”

This was not a recipe for fast completion of the design.

Ultimately, the specification was completed and Shima set to work on implementing the brand new design.

One key implementation resolution was that the Z8000 wouldn’t make use of microcode. Microcode would have damaged down the Z8000’s directions right into a collection of less complicated directions, hidden from the skin world, which the processor would execute. As a substitute, all directions could be ‘hard-wired’ into the logic of the CPU, a more difficult method for the designer. Shima would later focus on how way more tough the Z8000 was to create when in comparison with the Z80, and the way it examined the boundaries of the instruments that he had obtainable:

“… there are such a lot of directions in Z8000 it’s unattainable to retailer all of check vector for debugging within the reminiscence of check bench anymore. Additionally, MOS course of was getting denser and denser and in addition the scale of the defects in masks was getting smaller. That’s it was not really easy to seek out the absolutely purposeful die.”

The advantage of this method was that the Z8000 wanted fewer transistors than would have been the case had it used microcode. This meant that it ought to have been cheaper to supply. In idea, it additionally ought to have given the Z8000 efficiency good points over a microcoded competitor.

The extra complexity of this method, although, delayed the Z8000’s completion. It could take Shima about two years to finish the designs of the Z8001 and Z8002.

Shima was nonetheless engaged on the implementation of the Z8000 in June 1978 when Intel launched the 8086, its first 16-bit microprocessor. Faggin instantly knew that he had missed one among his key targets, to be first to market with a 16-bit microprocessor. Zilog now needed to play catch-up.

In some ways, the 8086 was much like the Z8000. It was designed with a segmented reminiscence structure and would allow automated translation of 8080 meeting code into directions that may very well be utilized by the 8086.

There have been some key contrasts with the Z8000, although. Intel had gone for a much less bold structure, one which regarded like an improve on their 8-bit 8080 quite than a contemporary method. The Z8000 had sixteen 16-bit ‘general-purpose’ registers and by comparability the 8086 had a a lot smaller set of registers and supplied a lot much less flexibility in how they had been used. The 8086 was additionally restricted to 1 MB of reminiscence, in distinction to the 8 MB of the Z8000.

Intel had additionally selected an implementation that used microcode. The 8086 wanted extra transistors (greater than 50% greater than the Z8000) for a much less spectacular design.

However Intel’s superior manufacturing enabled it to compete even with this penalty. And the microcoded design helped to allow Intel to get the 8086 to market first. It had taken solely a little bit greater than 2 years from the inception of the 8086 to it happening sale in June 1978.

The primary working silicon of the Z8000 CPUs solely turned obtainable in early 1979, over six months after the 8086.

And the Z8010 Reminiscence Administration Unit wasn’t obtainable till even later. It could take nearly a yr after transport the Z8000 CPUs for this significant element to ship. Till then, clients both needed to make do with both a 64 KB system, or to improvise reminiscence administration utilizing their very own ‘discrete logic’ designs which might be each pricier and supply fewer options.

Against this, they may purchase the 8086 which included the performance to handle 1 MB in a single chip. Intel now had an 18-month lead for many sensible functions.

However all was not but misplaced. Zilog was backed by the sources of Exxon began to market the Z8000 with the assistance of a ‘Tremendous Hero’.

They created a collection of comedian books showcasing the capabilities of the brand new design. The character referred to as ‘Captain Zilog’ led the best way in displaying off the Z8000. Captain Zilog battles ‘Diabolicus’ who controls the ‘largest and most advanced computer systems within the universe’ (and in addition apparently has ‘a little bit mail order pc restore faculty on the aspect in Santa Clara’ – presumably an in-joke that has been misplaced within the passage of time) however isn’t any match for Captain Zilog and his Z8000.

Here’s a pattern of a few of the dialog, which supplies a way of how the deserves of the Z8000 had been being marketed:

THE Z8000 HAS A REGULAR ARCHlTECTURE! THAT BIG COMPUTER DESIGN FEATURE AVOIDS THE BOTTLENECKS INHERENT I N TRADITIONAL MICROPROCESSOR DESIGNS!

WlTH 418 OP CODES, THE Z8000 PERMITS COMPACT PROGRAMS! THROUGH-PUT RIVALS THAT OF LARGE COMPUTERS!

THERE’S 32-BIT OPERATION CAPABILITY BUILT INTO THE Z8000 THAT INCLUDES MULTIPLY AND DIVIDE!

Captain Zilog obtained the Z8000 seen, however would it not be sufficient to take again the benefit from Intel?

In September 1979, Motorola introduced the 68000. The Motorola crew, sensing that that they might lose in the event that they launched with a design that was near Intel’s, aimed to leap forward with a way more superior design.

Crucially, not like the 8086 and the Z8000, the 68000 prevented a segmented reminiscence structure in order that it provided a steady reminiscence area of as much as 16 MB, a key benefit for customers trying to implement extra advanced or demanding purposes.

Intel needed to reply. Recognising that the 8086’s structure was weak when in comparison with its opponents, Intel centered on its strengths: specifically the vary of peripheral chips that had been obtainable for the 8086, which far outnumbered the vary obtainable for the Z8000 or the 68000. Quoting Intel’s Invoice Davidow the 8086:

“It turned out that we solely had one choice that was to promote at a system degree as a result of we had been worse on the element degree, however we had all of those different issues going. I keep in mind considering to myself, if we are able to place ourselves round the truth that that is crucial technical resolution the company we’re promoting to will make for the subsequent decade, we are able to speak to individuals at a really, very excessive degree, and speak to them concerning the quantity of engineering time concerned, how complete system price will likely be much less as a result of we could have all these peripherals, how you’ll write your software program sooner, we may focus the promoting processes in our favor.”

The mission to promote the 8086 turned often called ‘Operation Crush’. It was accompanied by a advertising marketing campaign, designed by Regis McKenna – who would additionally work with Steve Jobs at Apple – wherein Intel spent over $2 million.

Crush labored. Intel set a goal of two,000 design wins, and it will definitely achieved properly over that quantity. For programs that wanted probably the most processing energy, the Motorola 68000, with its non-segmented structure, was very enticing, however for different programs the 8086 received. Zilog had an early win with Olivetti for a desktop pc system, the M20, however there was little else of notice.

In 1979, the 8086 was joined by the 8088, which changed the 8086’s 16-bit information bus with an 8-bit one. The 8088 was slower than the 8086, nevertheless it was even cheaper to make use of.

After which in 1980 IBM selected the 8088 for the IBM PC. Furthermore, the IBM PC was an open system – made out of elements and working software program designed by different companies – in order that others may construct shut or (ultimately) absolutely appropriate variations.

Faggin would later say that he believed that IBM chosen Intel quite than Zilog due to Exxon’s possession of Zilog. Exxon had ambitions to be a serious IBM competitor in ‘Info Expertise’ and Zilog was a key a part of that ambition. Choosing the Z8000 would have meant counting on a robust up-and-coming competitor.

It was recreation over for the Z8000. Captain Zilog had been ‘Crushed’.

Shima returned to Japan firstly of 1980, fulfilling a promise to his spouse that their youngsters could be educated in Japan. Faggin left Zilog on the finish of 1980. Zilog would by no means problem Intel within the microprocessor market once more.

So why did Zilog fail when, in lots of respects, it had a aggressive and arguably higher product than Intel? Zilog was owned by Exxon, who was each bold and in addition comfortable to bathe massive quantities of money on its promising subsidiary.

These would possibly appear to be benefits, however they got here with vital drawbacks. Zilog was inspired by its father or mother to do greater than is likely to be affordable for a startup of its measurement. So while the money was obtainable to assist improvement, the sources inside Zilog had been unfold skinny.

Then there was the Z8000 itself. It fell between two stools. It wasn’t the most affordable, nor did it have probably the most superior structure. It was destined to be second selection in any contest.

Moreover, it’s additionally controversial that the Z8000 was centered on the mistaken targets. Code measurement was a key goal, however the extra reminiscence that was obtainable with 16-bit programs made this much less vital.

Zilog saved the price of the Z8000 itself down utilizing fewer pins, reminiscence segmentation and avoiding microcode, however the CPU itself could be a small a part of the price of a whole system. The extra MMU chip elevated prices. And reminiscence segmentation meant that the Z8000 would lose out towards the Motorola 68000 when it got here to selecting a CPU for extra highly effective Graphical Methods equivalent to Apple’s Lisa and Macintosh.

After which there was Crush and IBM. Intel’s advertising marketing campaign was vastly profitable as a result of it centered on the problems that mattered to potential clients and bought a whole answer. Zilog had, for a very long time, had a weaker and incomplete providing.

Did Zilog have it too straightforward with the Z80? Bernard Peuto would say that the Z80 had bought itself. Intel had struggled with the 8080’s successor and knew that they needed to deal with making the 8086 successful.

So it was most likely recreation over for the Z8000, even earlier than IBM chosen Intel’s 8088.

Zilog had made extraordinary progress within the few quick years after its founding. Quoting Federico Faggin:

“… I feel really that for 5 years Zilog set the tempo for the business. Zilog was for its first 5 years of life, possibly 4 you’ll be able to argue, the pacesetter within the business, what Intel later turned.”

However in the long run, even Captain Zilog couldn’t defeat Large Blue.

For extra commentary on the story of the Z8000, notes on sources, evaluation of the Z8000’s structure, and the way the Z8000 lives on in the present day then please turn into a paid subscriber.