Back in ancient times (1997), there used to be just one CPU product line. For instance, Intel would sell Pentiums until they got to around $100.
With 1998 and the Celeron, the idea of two product lines came into being: the regular line and the budget line. Both Intel and AMD figuratively did a kneecap job on their budget processors (by reducing or disabling cache), then sold them at much lower prices.
After about four years of this, though, this strategy seems to be going by the boards. Rather than have two product lines, and have CPUs in the upper division just go away at a certain (relatively high) price point, just let the old processor get older and reclassify it as a budget product.
AMD basically announced this strategy a few months ago. Durons will be gone by the end of the year. Replacing them will be first Palomino and/or Thoroughbred XPs, then Bartons, then Clawhammers.
What may not have been so obvious is that Intel is less publicly doing much the same thing lately.
The initial Tualatin PIIIs turned into Tualatin Celerons. After a short trial with 128K Willies, Celerons are now essentially the old Willies.
This makes bottom-feeding more appealing, and makes this strategy more appealing.
Old CPU, New Mobo
When a lot of people buy a new computer, they often ask “What’s the upgrade path?” and often, there just hasn’t been much of one.
As we reach a point where even the cheapest processor will pretty much get the job done, though, the idea of low-balling the CPU becomes more appealing.
The ideal time to do something like this is after the next generation of processors/mobos (which will support the cheapy processor) exists. The more expensive the newbie CPU is, the better this works.
So in the case of AMD and socket A, once Barton comes out, that will be the time to pick up a cheap Athlon and the most advanced mobo at that time. By that time, it could well be a $50 Thoroughbred and a KT400 board. A year later, you stick in a $50 Barton, and I suspect the owner will be set for some time to come.
In early 2004, we’re probably looking at much improved .09 micron Sons-Of-Clawhammers with prices to match. What you do is to buy the latest mobo and the original Clawhammer (making sure that the old Clawhammer works in the new mobo, of course). You can upgrade to Son-Of-Clawhammer later.
On the Intel side, if sometime late in 2003, you decide to migrate to a PIV platform, once a nice Prescott-supporting Springdale mobo, (maybe if you’re lucky a DDR-II supporting mobo will be around by then), you buy the cheapest processor that will run such a system. Tomorrow’s Northwood will probably become the year-after-that’s Celeron.
Why This Makes More Sense Now
This is obviously not a bleeding-, cutting-, or even dull scissor-edge strategy, but as the argument for higher and higher clock speeds becomes progressively more absurd for the average person, it makes more and more sense.
We’ve presented the same idea before in the form of “rent-a-CPU,” but that was more a temporary tactic for overclockers. This is more a strategy for Joe Sixpack, particularly cheap Joe Sixpacks.
The market is again shifting towards the low end. AMD might find itself selling more Durons than Athlons next quarter. Intel’s Celeron sales are chewing into PIV sales. So people are already voting that way with their wallets.
The problem with that is that they’re getting a somewhat lower price upfront, but the average budget system comes with budget components, and usually, you lose a lot more capability than you save money with such systems, both in immediate performance, and later longevity.
I suspect it will be easier to sell at least some Joes on this notion when you can tell him that he really isn’t buying a budget, reduced-capability processor anymore, but rather one that was a top-of-the-line $300-$500 processor twelve-to-eighteen months ago.
Plus, do not underestimate the trauma Joe goes through at the idea of getting a whole new machine and getting all his stuff from old to new. This is especially important when you’re going to be the one who’s going to end up doing it for Joe. 🙂
Telling Joe that a 15-minute CPU upgrade can postpone that for him for a few years can be very, very persuasive.
The Last Upgrade For A While?
I don’t know about you, but for the average person, once you get him up to current speed, I have real problems imagining how future advances are really going to help Joe very much doing what he does on a computer for the next few years.
Forget recession, or stock market crashes or 911. This is the real cancer sapping the strength of the desktop industry.
We live in an enclave in the overall computer world. It may be due to gaming or youthful competition, or hobbying, or all of the above, but “Why more?” is getting asked more and more.
Actually, we even hear it here, though it’s usually in the form of “I’ll give up MHz for less fan noise.” I think fewer people would be saying that less often if MHz really mattered more.
It’s not like there aren’t things that could be done. It’s just that nothing that would appeal to everyone look like they’re going to be done anytime soon.
If that doesn’t happen, and fairly soon, at least in the U.S., what are considered to be bad days for the computer industry will look like the good old days a few years from now.