Paul,
Tony - I think THIS says it ALL:
"While there are, of course, alternatives to Intel motherboards, some dealers and customers specify all-Intel machines, putting pressure on the distribution channel. µ"
This says it some more, from M. Magee:
Before we begin our examination on how Intel will shape up on the microprocessor front this year, we'd like to say that this isn't just our opinion. We've talked to a number of motherboard manufacturers and other Intel partners since the beginning of 2002 who like what they see on the roadmap. Typically, such manufacturers and OEMs do not find themselves in a particular microprocessor camp - many of them sell products which include AMD and Via products too.
The jury is still out about AMD on this one:
Although there were recently some questions over the efficiency of Intel's move from .18 micron process technology to .13 micron process technology, as far as we can tell from the outside, the change seems to have progressed pretty satisfactorily.
Good segmenting helps, and Intel has been listening to the customer:
Instead, we think that the way Intel has segmented the CPU marketplace this year, and its introduction of products across a wide range of platforms, puts it in a strong position to achieve its targets and please its shareholders.
We also think that Intel, over the last 12 to 18 months, has done much to listen to the plaints of its distributor and system integrator partners, although in the latter case, it still has much to do.
The big OEMs have a lot invested with Intel server CPUs, and won't make a willy nilly switch over to AMD anytime soon:
Server CPUs
We'll put aside the Itanium for a while. Intel has had the X86 server market sewn up for several years now with its Xeon processors, which are pervasive in big business and have a wide backing from the usual suspects including Compaq, HP, IBM and Dell.
These firms have expertise in supporting server chips, which go through a completely different testing and life cycle than desktop CPUs because of the applications they run on and the type of functions they serve.
Intel gets it right in 2002
Barring major cockups, SNAFUs
By Mike Magee, 11/02/2002 11:46:20 BST
INTEL MAY HAVE GOT it pretty much right in 2002, although this conclusion is subject to a few caveats, explored below.
After a truly terrible 2001 and an equally horrible 2000, Intel seems to have finally got its act together, judging from roadmaps we saw in the last quarter and the latest ones we saw just a week ago.
Could any of this be anything to do with the fact that Intel's Paul "Steely" Otellini has become more involved with making the company, as he might say, execute more efficiently? Before we begin our examination on how Intel will shape up on the microprocessor front this year, we'd like to say that this isn't just our opinion. We've talked to a number of motherboard manufacturers and other Intel partners since the beginning of 2002 who like what they see on the roadmap. Typically, such manufacturers and OEMs do not find themselves in a particular microprocessor camp - many of them sell products which include AMD and Via products too.
Nor would it be fair to look at these roadmaps without saying that much of it depends on Intel being able to manufacture enough products at the .13 micron process to satisfy the demand for such chips. Although there were recently some questions over the efficiency of Intel's move from .18 micron process technology to .13 micron process technology, as far as we can tell from the outside, the change seems to have progressed pretty satisfactorily.
We are not concerned either, here, with whether the performance of Intel CPU at the top end are better than its main processor rival, AMD. Such questions obviously have an effect on sales, particularly for the gaming, enthusiast and overclocking users who seem to have drawn the right conclusion on price performance.
Instead, we think that the way Intel has segmented the CPU marketplace this year, and its introduction of products across a wide range of platforms, puts it in a strong position to achieve its targets and please its shareholders.
We also think that Intel, over the last 12 to 18 months, has done much to listen to the plaints of its distributor and system integrator partners, although in the latter case, it still has much to do.
Desktop CPUs
One reason we think that Intel has managed the change to .13 micron for its desktop CPUs is that the last three to four roadmaps we've seen show a level of consistency, with the Pentium III fading rapidly into the background. The move to using the Pentium 4 core in Intel's Celeron offerings is also an indication that this move is successful.
From the sales point of view, Intel's decision to phase the 533MHz front side bus in May without making any price adjustments is also an indication that it is reasonably confident about its ability to execute and move from the Willamette core to the Northwood core. The segmentation of clock speeds also appears to be a rational move from the standpoint of sales, but Intel needs to be careful that the market doesn't get overconfused with Pentium 4 chipsets. Perhaps it has realised that itself, so postponing introductions we expected.
Notebook CPUs
We're very much more puzzled by the notebook roadmaps, although the segmentation, particularly on the Pentium III-M side into low and ultra low voltages makes sense to us.
The puzzling aspect on the notebook CPU front is that Intel appears to want to sell its new Pentium 4Ms as performance and other processors, while still maintaining a strong Pentium IIIM offering across performance and other sectors.
We note that Intel is being very coy about what it proposes to do on pricing when it introduces the Pentium 4Ms in April. Does it have a cunning plan up its sleeve to slash prices on Pentium III-Ms? If not, we do not really understand how it can keep both roadmaps going simultaneously throughout the year.
One possible reason is that it is possibly hedging its bets on public acceptance of the Pentium 4M.
Despite Intel's insistence that its SpeedStep technology and the better thermals of its notebook processors make its processors compelling, we are still hearing that large OEMs, Dell in particular, are still not able to implement this kind of technology properly. This needs fixing.
As we pointed out in a separate article last week, companies introducing "desktop notebooks" could also have an impact on sales of Intel's Pentium 4Ms. The .13 micron desktop CPUs are fast and run fairly cool in a notebook chassis.
Server CPUs
We'll put aside the Itanium for a while. Intel has had the X86 server market sewn up for several years now with its Xeon processors, which are pervasive in big business and have a wide backing from the usual suspects including Compaq, HP, IBM and Dell.
These firms have expertise in supporting server chips, which go through a completely different testing and life cycle than desktop CPUs because of the applications they run on and the type of functions they serve.
We think the way Intel has segmented the server market this year is quite interesting and will probably pay dividends.
The introduction of processors for the so-called "blade" market, pedestal servers for work-a-day functions, rack servers and enterprise servers will all receive wide industry backing from HP et al during the course of this year.
AMD has only comparatively recently put in the infrastructure to support its dual Athlon MP processors, and so we'd expect products based on these to have little impact for at least the first half of this year.
Similarly, although Clawhammer and Sledgehammer appear to be compelling processors, realistically, AMD cannot make big gains from these chips until 2003 at the earliest.
The Itanium now appears to be a horse running on a different racecourse to us. It's pretty inconceivable really that after spending god knows how many billions of dollars, and getting its ISVs and other vendors to commit to the platform, that it's just going to shunt it out of the way any day now for a processor based on the X86-64 architecture.
Nevertheless, this year will be something of a watershed for the Itanium platform. When Intel formally launched the chip in May last year, we didn't have high expectations that it would sell many and the major market research companies appear to have confirmed that's the case.
This year looks like a re-run for the Itanium platform with the introduction of McKinleys. If Intel doesn't manage to sell many of those this year, then maybe it will have to seriously rethink its strategy. µ
See Also
Latest Intel roadmaps
theinquirer.net
This one is worth saving to throw back at the "Intel has been farking up a lot lately" type posts.
Tony |
