Intel seems to have solved part of their production and/or packaging issues for the Coppermine processors, however availability of 800MHz+ parts still seems to be very limited. Several resellers I spoke with indicated that they have no trouble finding parts up to 733MHz, but above that they are mostly available only through the gray market – or not at all.
Even though Intel has been maintaining that the processor shortages have been due to ‘strong demand’, there is evidence that this is actually not the entire reason. There have been several reports of an Arizona fab that was left unfinished last year, and Intel has admitted having some packaging issues with the new FC-PGA package. Interestingly, just this week, Intel announce that they would spend almost $5 billion this year on chip manufacturing capacity, claiming that increased global demand may well create a semiconductor shortage in the near future.
There is some truth to this, as the growth in pagers, cell phones, PDAs and other portable electronic devices has skyrocketed in the past year. One area of concern is that capacitors, resistors and other discrete components are becoming more and more scarce as demand is outstripping supply. It is unknown from the news reports whether Intel intends to invest in the manufacturing of these types of devices.
Slot 1 appears to be all but a thing of the past, at this time. Several motherboard manufacturers told me that they will produce no new Slot 1 motherboards, at the suggestion of Intel. Supplies of these processors will likely all but dry up by early summer, if the information I have received is correct.
A ‘new’ Celeron chip, based upon the PIII Coppermine core, was introduced this month to a somewhat lukewarm reception. The main improvements over the ‘old’ Celeron is the 256-bit wide L2 cache and SSE instruction set. The bus speed remains at 66MHz, and the cache size remains at 128KB. Performance, in other words, doesn’t improve much. In addition, since the ‘full blown’ Coppermine is still having problems reaching 800MHz+ speeds, these Celerons are not likely to be an overclocker’s dream.
Though AMD has been shipping some 1GHz Athlons (K-75) during the month, they are not exactly what would be called ‘readily available’. Apparently, Gateway is taking the lion’s share of whatever is produced, with everyone else having to take the leftovers. Part of the reason for this is yields, of course but another reason is very likely marketing.
With the Thunderbird processor (based upon the .18 micron Copper process out of Fab 30), speeds are expected to start at or near 1GHz and ramp from there. It would make little sense to offer the same speeds from both models, as the Thunderbird is likely to vastly outperform it’s K-75 cousin at similar clock rates. For this reason, the K-75 processors will be worth much less in the marketplace, and it would not be the best marketing strategy to be left with a large inventory of 1GHz processors that have to be sold at a discount.
There has been some speculation as to why K-75 prices have dropped recently, considering the fact that demand is up, and Coppermine processors are scarce. The reason for this is very likely the same one given above – room must be made for the Thunderbird processors. There is always an issue when pricing products, with the ‘sweet spot’ (between $300 and $500) reserved for the speeds intended for the ‘mainstream’ user.
Of course, the Spitfire also comes into play here, as it is intended for the ‘value’ segment, where the Celeron processor plays. Speeds on this chip are not expected to exceed 700MHz (perhaps a bit higher) in the short term, partly because there is little reason to do so (nothing faster than that expected from Intel) and partly for the same marketing reasons mentioned above. If the Thunderbird is playing at the high end (1GHz and beyond), and the Spitfire plays only at the low-end (700MHz and below), this leaves room for the K-75 to play in (700MHz to 900MHz+, most likely).
While AMD has been saying that both Thunderbird and Spitfire will be available in Q2, it seems that they mean late Q2 – or, near the end of June. Thunderbird is scheduled to be released first, with Spitfire debuting shortly thereafter. It is probably wise to plan for the fact that these processors will not be readily available until July, at the earliest.
AMD is being very tight-lipped about the cache sizes and organization of these chips. Though this has been determined internally already, AMD apparently feels that it is to their benefit to keep the information secret, just as they did with the original Athlon specs/performance the same time last year. In that case, they were able to catch Intel flat-footed, and perhaps hope to do the same thing again.
Though AMD hasn’t actually said it, there is one thing we can surmise about the Spitfire based upon the chipset it will utilize – at least some of the speeds will very likely operate on a 133MHz FSB. This alone should make it more desireable than the equivalent speed Celeron.
Strangely, it used to be Intel that would keep all of their designs secret, while AMD would try to trumpet their ‘future advancements’, but now we are seeing exactly the opposite. Who’da thunk?
The current market for Athlons appears to be quite good. One reseller I spoke with indicated that 90% of their business is Athlon based, even with Pentium III processors having better availability. In addition, there were rumors that Dell is talking about a small quantity of Spitfire processors (about 200,000). According to my source, Dell is not happy with the Intel chipset roadmap, which appears to be faltering right now, and is concerned about the problems with Coppermine availability. The word right now is that Acer will be contracting to build Dell notebooks and motherboards, and that Q3 could see orders for Spitfire boards. The more skeptical, (including myself and AMD) believe that this may be a ploy to put pressure on Intel to deliver the goods.VIA (Cyrix/IDT) Update
VIA announced this past month that they would be spinning off the Cyrix division into a separate company, which will focus mainly on the Asian market at the very low-end. According to sources close to VIA, the IDT core is still intended for their integrated chipset that will compete with Intel’s Timna chipset at the low end in other markets.
Be the first to discuss this article!