Re: More and more I see the K7 as a niche product
I think your perspective is colored by your conviction that there will be no volume out of Dresden for at least a year - and you may be right. But if Dresden were to be successful by Q1 or even Q2, then AMD would have an excellent chance of maintaining a meaningful MHZ and performance lead over Intel until early 2001, maybe a lot longer. As has been demonstrated by what Intel did to AMD this year, commanding the high ground in x86 processors can be extremely lucrative, and being stuck in the lower 2 thirds of that market can be disasterous.
A niche product? Apple is limited because it isn't IBM compatible. If AMD ships volume out of Dresden at substantially higher clocks than Intel, the market will switch in a heartbeat - look what happened to market share when AMD was competitive in price and speed late last year.
Coppermine in portables is going to be iffy until the Rambus thing is straightened out - they've made Rambus a big feature of the Coppermine family and Rambus uses too much power. Coppermine without rambus may not sell well if a big PR push has been made for the benefits of Coppermine+Rambus. Maybe some news will come out of IDF that some versions of Rambus can use as little power as SDRAM.
Meanwhile AMD has released a 500MHZ K6-2 and reports that they'll begin making copper K6-3s. The K6 core is competitive with Coppermine in integer, floating point is of less importance in notebooks, and a copper K6-3 may dissipate substatially less power than an Aluminum Coppermine. AMD would make additional inroads into the portable market.
If AMD enjoys any success at copper .18 they'll be ahead when they have to go to copper .13, just when Intel is facing it's challenge.
I'm also encouraged by the new AMD embedded chip. This was the market that AMD did so well in with the 29K, and it may help explain Microsoft's enthusiasm for Athlon. Intel bought ARM, which, if successfully developed, would leave Windows with no compatibility benefits when non PC devices like handhelds are being designed. If AMD's new chip becomes dominant in this area, many developers will find it easier to stick with Microsoft technology when they move to embedded devices, since many of the same tools would work in both areas. So we may continue to see Microsoft push hard to help AMD succeed.
Just one opinion,
Dan |
