Intel has announced it’s scrapping its 4 GHz P4. That’s a big turnaround.Intel got where it is today by cranking the megahertz, and then the gigahertz, just as high as it could and as quickly as it could, hoping competitors wouldn’t be able to keep up, and trumpeting clock speed as the only thing that really mattered.
When it designed the P4, it extended its pipeline to ridiculously long lengths, allowing it to pump up the clock rate, but the efficiency was so low that Intel had to be ashamed of it. The last of the P3s cleaned the P4’s clock. As did a number of AMD’s chips.
Now Intel is having difficulty reaching 4 Ghz. AMD still has room to ramp up its speeds, but it hasn’t even reached 3 GHz yet. They’ve been taking other approaches to increasing speed.
Now Intel’s taking yet another page from AMD’s book. First, Intel clones AMD’s 64-bit instruction set, next, Intel replaces clock speed with model numbers, and now it throws in the towel on the gigahertz race.
It’ll be interesting to see how Intel’s marketing adjusts. And while I don’t expect AMD to topple them any time soon, if ever, it’ll be interesting to see if AMD manages to turn this into another opportunity.
The terrible thing is, when they switch to multiple-core CPU chips, high-end applications like Oracle will charge license fees as if there were two (or more) separate processors. Thank goodness for open-source alternatives.
That is a big, legitimate problem. I never really understood licensing on a per-processor basis. It’s not like software wears out faster when you use it harder, like a car on lease does. It’s more like a store charging me $10 for a telephone because I live alone but charging a 4-family household $30 for the same phone because they might use it more.
Hopefully enough people will complain to get the software makers to abandon per-processor licensing. After all, within a couple of years it won’t be possible to buy what used to be a single-processor system. If they need more money, then they need to either raise their prices or figure out how to expand their market. Either that, or, as you say, MySQL and company are going to see their usage jump considerably.