History repeats itself. Way back in the day Intel made chips with model numbers like 8086, 80286, 80386 and 80486. Since the first two numbers were the same, they decided to drop them and name the chips the 286, the 386 and the 486. They placed a math coprocessor on the 386, making it faster but more expensive. They did this for the 486 as well. Then they found that people wanted lower priced chips, so they unveiled the 386/486SX (which somehow stood for “single precision”) and the 386/486DX (“double precision”). The SX lines were literally identical to the DX lines, but the math coprocessor was broken (literally physically broken). They would then happily sell you a math coprocessor chip – to replace the one they broke for you. Sneaky? Yes, but it worked – if you weren’t concerned with speed you could save some money and then upgrade later.
When the 286 came out demand was so high Intel farmed out some production of it to a startup chip maker named AMD. After the demand settled they told AMD their services were no longer needed. Imagine Intel’s surprise when AMD started coming out with chips whose architecture mimmicked the x86 architecture. Obviously being privvy to special Intel documents gave them the knowledge needed to be able to make these cloned chips. Interestingly enough their case held up in court – the judge figured they could have made the cloned chips even without the Intel knowledge (it just might have taken longer).
AMD even named their chips the 386, 486, etc., for which they were sued again. Intel claimed they owned the copyrights for those names, but of course they didn’t have a case – you can’t very well copyright a number. Intel quickly tried to rename their chips the i386 abd i486 (“i” for Intel – this was long before the iMac) but those names never took (and wouldn’t have helped them in court anyway). Intel made the argument that the numbers were conjured from thin air and that the sequential-ness of them all was coincidence. They also made the argument that SX and DX were similarly conjured from thin air, that they didn’t stand for anything, and that the DX2 was not a “clock doubled double precision” chip. This went so far that the clock tripled 486DX chip went under the name 486DX4, not the logical 486DX3, so that they could help the name argument (and make a quick buck off those who assumed it was a quadruple speed chip).
It didn’t work – what Intel needed to do was to put their money where their mouth was and conjure up a new name for the 586, so they dubbed it Pentium – a made up word whose root was “penta-” (five). This worked – Pentium took off as a strong brand name. AMD’s 586 chip, which they dubbed the K5, looked puny in comparison. Even their quick follow-up, the K6, bombed.
AMD’s fallacy was that their chips were never as fast as Intel chips, even at the same clock speeds. This was less relevant, since AMD entries never debuted at a clock speed as high as their Intel counterparts.
Intel, meanwhile, unveiled the Pentium successor, Pentium II. They had pretty good ties to the strong “Pentium” brand, and since the greek prefix for six was “sexta-” (a “Sextium” wouldn’t have worked, they sumised), and since sequential naming got them in trouble in the past, they kept the Pentium name, following up the Pentium II with the Pentium III and Pentium IV. They also came out with another “budget” chip, the Celeron (whose name became originally became synonymous with “celery”) which was a Pentium II without cache. This was in response to a fad notion of the Network Computer – that people en masse would give up their desires to own pricey machines and instead all make do with dumb terminals. Woefully slow, they eventually gave it a limited cache and it became an attractive option for low priced computing.
Then just after the introduction of the Pentium III, AMD finally released their Athlon chip (what would have been the K7). Finally, they released a chip with a higher clock speed than an Intel chip and one who could benchmark faster as well. In addition to being marginally faster on real-world applications, it was also less expensive. Ever since then Intel and AMD have been waging a war on clock speed and price.
The other half of hating Microsoft is hating Intel, since the “Wintel” architecture is the market beast. Therefore, those who love to hate Microsoft love to love AMD and their Athlon. Me personally I went with Intel’s Pentium III two years ago when I made my system, since Athlon was untested in the market. The added cost of locating an Athlon motherboard (at the time) negated any price difference the chip provided. And I think that while Athlon has proved itself (more or less) in the marketplace, I’ll probably stick with Intel, but as my next processor purchase is down the road some, I’ll keep and eye out on both.
One of the things that would happen with the Microsoft/Intel cozy alliance was that Microsoft would find something that they thought it would be great if the processor could perform as an operation, so they would phone up Intel and suggest it. Intel would agree and place it in their next chip. Now Microsoft is happy since they have software that uses the new operation (and they are the only ones so far that know how to use it) and Intel would be happy since now the most popular software code in the world ran better on their chips. This is why in the end they let AMD do whatever without further litigation – they just figured that they could do one better next time and AMD would just keep playing catchup.
Now history is repeating itself in a few ways. The Register is reporting that AMD is naming their next athlon the Athlon XP and had curiously delayed it to hit at the same time as Windows XP. They’re also doing away with megahertz as a method of naming chips. the 1.3GHz model will be named the Athlon XP 1500+, the 1.4GHz model is named the Athlon XP 1600+, and so on. This gives the consumer the illusion of added speed, and the name change gets them all chummy with Microsoft, who probably doesn’t have a problem burning bridges with Intel. Will it work? Who knows – the economy is going to shit as we speak (both before and because of the WTC incident) and so clinging to someone who you think is going to come out alive isn’t unheard of, but this does put another fun wrinkle on which chip I should go for.