I don’t care what Intel says, these new processors are spiffied up, die-shrunk Pentium Ms. Just like Hammers are basically Athlon XPs with a memory controller, these chips are Pentium Ms with desktop features attached.
The only people who are going to be impressed by the claims on power consumption will be those who are ignorant of what Pentium Ms can do today.
Not that this is a bad thing, it’s just not a great thing for overclockers. From the estimates and guesstimates floating about, on the desktop, these chips may be able to run 20% faster than current overclocked Dothans. Unfortunately for Intel, that’s not likely to get them significantly faster than current Hammers, much less future 65nm versions.
The real significance of these chips is the direction in which these chips are headed. Intel sees an increasingly mobile world where people care less about an extra 10-20% of performance, and more about battery life and having a space heater in the office or on the (server) farm.
These chips aren’t meant to level the playing field, but change both it and the game being played on it.
Hit ‘Em Where They Ain’t
If you look at the AMD lineup, it’s easy to see that their Achilles heel is the lack of a low-powered CPU stable that can compete with the Pentium M.
This is one game Intel is winning, so you can see why they’d like to turn it into the game. AMD has nothing like the Pentium M, so Intel would love to turn the CPU wars into a battle where they’re strongest and AMD is weakest.
Lower power can be leveraged a lot of ways. Think MiniMac. Think MiniMac a lot. You don’t want the constraints of a MiniMac, fine, it’s not meant for you. It’s meant for the Joe (and Josephine) Sixpacks who hate big ugly boxes, as well as for the Joe Suits with blade servers, and everyone looking for a serious computer in his or her phone.
If you read between the lines, Intel apparently thinks it can do more at 45nm. Their claim that they can eventually improve performance/watt by a factor of 10X accords well with rumors that leakage will no longer be a problem at 45nm.
If all this sounds directly contradictory to the high-performance/gamer ethos, you’re absolutely right. Intel will never say it, but they’re trying to evolve away from that, leaving those who are interested in raw power only behind as some sort of technical Neanderthals.
Is this making a virtue out of a necessity? No doubt. What’s important to note is that Intel is implicitly indicating that when push comes to shove, they’re going to choose the housewife over the raw power user, and leave AMD to fight the last war.
Consider this: since at least the age of the Athlon Thoroughbred, AMD has had a disproportionate share of those interested in high performance.
How much has this hurt Intel in the overall marketplace?
We have long believed, long before Intel developed a fever, that mainstream personal computing was going to turn onto itself and use technological innovations to make devices smaller, more portable, handier, cheaper. Intel’s actions are a big step in that direction.
And if that leaves folks like us behind, well, so be it. Let the dog wag the tail rather than the tail wag the dog. 90% of computer users will be quite happy with a little, cool, fast-enough-for-them computer.
It’s just that the other 10% are the people who read sites like this one. Well, the days of ruling the roost are drawing to a close. High-performance will become a niche.
Sort of like Mac users.
P.S.: Next– The Missing Links
It should be noted that these new processors represent a delay in new product introduction from Intel. These processors aren’t going to show up until the second half of 2006. Unfortunately for Intel, its 65nm manufacturing capability is supposed to start cranking up at the end of 2005.
In other words, Intel will be all dressed up with no place to go.
Well, just to get those brand new fabs up and running, Intel will have “missing link” CPUs. I’m not sure what else you can call products like Yonah or Dempsey. They’ll show up in early 2006, and offer relatively minor improvement over the current bunch, but after a short time in the spotlight, they’re supposed to jump aside for the real deal.
Intel said very little about them. Not much you can say about placeholders.
So Intel is going to start their 65nm fabs with one kind of processor, then start changing over to another six months later. Sounds like a mess to me.
Perhaps more importantly, why would you knowingly buy a “missing link” product if you were aware that a more advanced species is just months down the road? (Of course, for Intel, ignorance will prove to be bliss.)
I think it’s safe to say that these will be products you and your friends/relatives/associates ought to duck.