When Intel introduced its Atom processor, it included a spreadsheet of potential customers. The spreadsheet listed features: how many cores, how many threads, how fast the clock speed, and so on. But the spreadsheet also listed the features none of the potential buyers had.
The spreadsheet was incomplete. Intel’s customers would encounter new and unexpected problems. They would need new and unexpected solutions. Intel’s customers were smart, and they would try to solve their problems in ways that made sense to them. But Intel wouldn’t necessarily understand those ways. The company would have to learn by doing.
Intel’s customers included a lot of startups, and startups are in many ways a special case. A startup needs to develop its own culture. A startup has no customers, so it is its own market. A startup has to invent its own problems, so it has to invent its own solutions. When a startup chooses processors, it hopes to solve its own problems, but it might fail. And if it fails, it might reinvent the world.
Intel’s customers were not all startups, but they were almost all of them small. And the smallness was partly a matter of size. Intel’s customers were usually one person, with maybe one or two other people working with them. Intel understood this, and targeted its processors at the small end of the market.
All the same, Intel’s customers were usually small, but they were not usually dumb. They knew what core count meant, and they knew what clock speed meant. They knew that they had to plan ahead. They knew that Intel processors were fast, but they didn’t necessarily know that they were the fastest.
Intel’s customers were not always smart, but check the best mobile processors.
The smartphone processor market is changing fast as new processors come on the market. With so many processors to choose from, it’s important to select one that is a good value.
“Good value” is relative, of course. A $50,000 car is good value for a 5 percent discount, but almost worthless to everyone else. The price change affects different people in different ways.
The smartphone processor market is much the same. For consumers, a strong processor is very important. But so is battery life, screen quality, and user interface. In evaluating the processor, look at all three.
At the low end of the market, processors are relatively cheap, and it is common to buy more than one, so the purchase price of chips is not as important as the benefits each delivers. For users of smartphones with spending power below $400, the best processor will be the same one, whether it is $40 or $400.
For users at the high end of the market, a major decision is processor speed. Processor speed is the yardstick by which all smartphones are judged. The faster the processor, the better the performance.
The advantage of one processor over another is the value it delivers. (To be fair, processors can also deliver more battery, more screen, or more memory, but speed is the base measure.)
The processor market is complex. It has strong vertical markets, in which one company dominates, and strong horizontal markets, such as processors that differ only in the numbers of cores. Most people will make a processor choice by major feature. For new users, it is frequently best to pick a processor with as many cores as you can afford. If you have plenty of memory and want
The smartphone processor market is divided between a small handful of large players, and a large number of small players.
The large players include Qualcomm, Samsung, MediaTek, and Intel. These big companies design their own processors, so each has a distinctive look and set of features. The small players include Rockchip, Spreadtrum, and HiSilicon. These companies have either licensed the designs of the large players, or bought them outright.
The distinguishing feature of the big players is the size and quality of their engineering staffs. They have huge engineering departments, staffed by many thousands of people. With so many people, they can design and develop lots of processors at once. The smaller players do not have this luxury. They have to concentrate on one or two processor lines at a time.
As a result, they rely more heavily on the engineering talent of other companies. Qualcomm, for example, bought its processor designs from LSI (now Altera). MediaTek and Intel bought theirs from other chip companies.
So if a small player is not sure whether to hire a new processor designer, first ask around. Did someone here just quit? Did someone here get poached? Did someone here get a raise?
And if you hire, remember: Production is a series of small decisions, and the smallest of them is: “Who to build this with?”