that_crazy_doug,
By December, AMD will most likely have lost the MHZ crown, and will have to try and convince people that performance is made up of more then mhz while fighting the intel propaganda machine.
There are a couple of things working in AMD's favor. The cost of P4 systems will keep them out of the mainstream consumer market for several quarters. Also, the media seems to be catching on to the fact that P4 has low performance relative to clock. If P4 performance is disappointing, it is going to become a huge story in the press. AMD's ponies are going to compete very nicely.
The big opportunity for AMD is in the two processor server market. From the Register story:
Four way servers based on the ServerWorks chipset will jump from 700MHz PIII Xeon with 2Mb cache to 900MHz at the end of this year, while lowlier dual CPU servers move from 933MHz to 1GHz at the same time.
In Q1, AMD will have dual 1.5 Ghz Mustang servers. I suspect that these servers will match the performace of 4-way Intel Xeon servers. This level of performance will not be ignorable and is currently AMD's biggest opportunity.
I suspect that P4 in 0.18u will scale to around 2.2 to 2.4 Ghz. But, power and heat will be major issues. Combined with die size issues, P4 does not become viable until 0.13u -- or 2H, 2001. So, how will AMD be able to compete with a 3 Ghz P4 in, say, Q4 of 2001? My guess is that AMD will have a 0.13u (possibly SOI) 2+ Ghz Clawhammer in production by that time, and that it will perform favorable againts the speedy P4. Clawhammer die size will be much smaller than P4 and AMD will continue to be able to offer a price/performance advantage. Then we get Sledge with its dual core design and LDT that will provide 2, 4, 8, 16+ way high performance 64-bit servers.
Combine this with AMD's flash business and you have a powerful combination. I was getting a little worried about how AMD would fill their new flash fabs, but the strategic alliances with TI and LSI have removed this fear.
AMD is here to stay.
Pravin. |
