Updates to this story
I really find it truly remarkable that AMD has decided to replace Dirk Meyer – quite a man – with some glorified bean counter that came from Qimonda, for crying out loud.
Meyer, as we outline in a news story elsewhere on this virtual rag, headed up the team that produced the DEC Alpha microprocessor – a chip that's still some way ahead of his time. And in his role at AMD, his team that worked on the Opteron turned round the company's fortunes and gave Intel a scare that it will never forget.
Reaction round the web to the news has been interesting. Don Clark, at the Wall Street Journal, managed to talk to both Jerry Sanders III and to Hector Ruiz – characters as similar as cheese and chalk.
Sanders told the WSJ that disposing of Meyer was a “disgrace”, while Ruiz confined himself to the rather bland statement that he found it surprising.
What's more surprising is that AMD is showing signs of turning itself round – it forecast a profit with gross margins of around 45 percent for its next financial quarter.
Insiders told Clark at the Journal that the board was upset that AMD wasn't moving fast enough to make waves in the server market and for crying out loud tablet computers. Of course AMD and Intel are like Tweedledum and Tweedledee, except they're more joined at the hip than those Lewis Carroll characters. If what the Journal says is true, we wouldn't be surprised if Paul Otellini, Intel's CEO, is hearing the same murmurs.
But the Academy of Beancounters really does not have an earthly about what you have to do if you are designing a microprocessor, and just how difficult it is. We don't mean the circuit design and stuff, but the sheer gamble an Intel and an AMD take in launching products and anticipating whether they'll be successful in 18 month's time. Or not. Heck, sometimes they're not. The Pentium 4 was a disaster and Intel put a brave face on it forever.
AMD never appears to learn from its lessons, while Intel is actually at its best when its back is against the wall. I can hardly begin to count the times over the last 20 years when AMD managed to aim very precisely and execute very one-pointedly on shooting itself in the foot.
The prime example was an Intel Developer Forum not long after AMD launched the Opteron in New York. I met the AMD execs and they asked me what I thought Intel was making of it all. I said what's clear to me is that you guys have an 18 month to two year advantage over Intel. At the same time, I thought to myself inwardly that AMD was going to rest on its laurels, like it's done before or since.