x86s Are Fast Because They Have Volume
As a result of the economics outlined above, x86s tend to have a very distinct advantage against other MPU families. Anne & Lynn Wheeler recently pointed out in comp.arch that VAX unit volumes have traditionally been quite low. The MicroVAX II, a very successful product, shipped 65,000 units over 1985-1987; by comparison over 1985-1987, there was roughly 14,000,000 IBM PCs and clone systems sold. Intel’s large x86 volumes mean that they can spend more time and effort optimizing their CPUs, and probably still have lower average costs.
If the unit volumes are inherently low, you have to get profit based on *system* value that yields unusually high margins, so that the system profit essentially subsidizes the use of such chips. This strategy works quite well for a system vendor when the market dynamics and customer switching costs allow high margins, i.e., IBM mainframes to this day, and VAXen in the 1980s. The first MIPS were designed using 2 VAX 11/780s, plus (later) an 8600, plus some Apollos; the VAXen seemed expensive, but they were what we needed, so we paid.
Mainframes and minicomputer systems companies thrived when the design style was to integrate a large number of lower-density components, with serious value-added in the design of CPUs from such components. For an example of this, look at a VAX 11/780 set of boards.
As microprocessors came in, and became competitive, systems companies struggled to adapt to the change in economics, and to maintain upward compatibility. Most systems companies had real problems with this. There were pervasive internal wars, especially in multi-division companies. Each faction would espouse one of the following strategies:
- We can do this ourselves, especially with new ECL (a faster, hotter and more expensive circuit design style) or even GaAs (Gallium Arsenide, a faster, more expensive and fragile semiconductor material). This view was common among those who worked on supercomputers, mainframes and minicomputers chasing mainframes.
- We can build CMOS microprocessors that are almost as fast as ECL, much cheaper and with better features, functions, and performance than commodity products. IBM, DEC, HP and later Sun would all pursue this strategy.
- We should put our money in system design and software, and buy microprocessors. Apollo, Convergent, Sun, SGI, Sequent, and various divisions of older companies tried this approach.
Of course, later there came:
- We should buy microprocessors, add value in systems design, and use Microsoft. This lead to IBM’s PC division, Compaq, etc.
- We should minimize engineering costs and focus on distribution i.e. Dell.
Many companies pursued one strategy too many. Most of the minicomputer vendors went out of business. IBM, DEC, and HP were among the few that actually had the expertise to do CMOS microprocessor technology, albeit not without internal wars. I was involved in dozens of these wars. One of the most amusing was when IBM friends asked me to come and participate in a mostly IBM-internal conference at T.J. Watson, circa 1992. What they really wanted, it turned out, was for somebody outside IBM politics (i.e., “safe”) to wave a red flag in front of the ECL mainframe folks, by showing a working 100Mhz 64-bit R4000.
Discuss (4 comments)