The Alpha 21x64 family, the Motorola 68K family, and the AMD 29000 family demonstrate that a high?quality architecture or superior performance does not guarantee commercial success. A company’s ability to control the market is far more important than a processor’s technical merits. Is this statement true?

Question

What will be an ideal response?

Answer 1

In principle, a processor with a higher performance than another should prevail (assuming comparable costs). In
principle. In reality, the situation is rather more complicated. Many factors are involved; for example, the scale
of production. An organization with a large fraction of the market can spread costs (design and development,
marketing, etc.) over a very wide customer base. Equally, a manufacturer with a much smaller customer base
has to spread costs less widely which means they must either reduce the cost of research, development,
manufacturing, and promotion, or they must charge a higher price for the product.
Similarly, there is the compatibility issue. If you design a new processor it has to be connected to a host system
which requires a motherboard. You then have to convince motherboard manufacturers to support your
product.
There is also the issue of software compatibility. If you make your processor compatible with mainstream IA64
processors at the machine?code level, you are trying to make a better Intel processor than Intel. If you change
the instruction set architecture, you have to convince Microsoft and those who create applications to produce
compatible versions of their software.
Nicholas FitzRoy?Dale makes the following interesting observation in an article on VLIW and EPIC processors,
“These products (i.e., early VLIW processors) failed commercially; the prevailing opinion seems to be due to their
position as startups. Thus, several small technical mistakes combined with the difficulty of selling expensive
machines from a poorly?established company, rather than any major architectural deficiencies, caused their
downfall.”