Jim,
If that is a real Willy, then we may have seen our first real Willy benchmark (11,483 3Dmarks). This is about 5% faster than what my back of the envelope calculation says an SDRAM based PC133 T-bird would get at 1.4 Ghz.
So by the time this is out, what will Mustang be able to do? I believe that Mustang will increase performance by at least 5% (given Scumbria's comments about T-bird's L2, I believe this is reasonable). We will also see speed improvements due to the 266 Mhz bus, DDR memory, and larger caches. My best guess is that the die of Willy is large enough that Mustang with 1MB of L2 cache will be about the same size as Willy.
Therefore, clock for clock, my best guess at this time is that Mustang will be 10-30% faster than a similarly clocked Willy. Since we all know that "Mhz sells (tm Jim McMannis)", we may wind up next year in a position where 1.5 Ghz Mustangs are the performance champs, but 1.7 Ghz Willys command top dollar.
If this happens, it really limits the prices AMD will get for their chips, and it will limit how high AMD's stock can go. Under those circumstances, I really don't see how AMD could go much higher than $400-$500 a share by the end of next year.
Mike
|
