Saturn V: I agree (based on comments from Haim Barad and others) that Athlon will not be able to directly benefit from code optimized for SSE. And you deserve credit for point 4 in your second reference post re SSE; that was a significant bit of foresight. Did you think they would need it as much as they did?
With your second assertion I have several quibbles. As a physicist I probably have even less insight into business than the engineers, nor do I think that benchmarks correlate strongly with business success. However, I think that Intel's action in changing the benchmark so that their numbers would be more impressive speaks volumes about the importance of benchmark numbers to certain segments of the market.
To some of your other points: It certainly does not seem from currently available (pricewatch) info that Athlon commands less of a premium than PIII. The imminent arrival of 700 MHz CUmine seems quite unlikely to provide the coup-de-grace to AMD, given that Athlon's floating-point speed advantage appears unlikely to go away once the benchmarking field is again leveled. (Will it be leveled? Well, most game houses already code for 3DNow, and a poster on JC's hardware site implied that it was quite simple for him to modify his code so that his Athlon's FP perf improved by ~40%. Seems doable.) As to the impossibility of AMD profits, the naivete of its investors, the hopelessness of its business model, and its general reasonableness as an investment: At the moment I would say the odds for a profitable Q1 2000 are better than 50-50; while hardly a ringing endorsement it would be a significant turnaround for AMD and its improved business model. That model, by-the-bye, is not the discount one you disparage, but the rather more compelling one of raking off some of that rich high-end desktop, enabled by the sound benchmark thumping by the Athlon, an edge it now seems it may be able to hold until Willamette. That is a significant window, not enough perhaps to build a long-term success, but probably enough to change professional money's opinion of valuation. And that of course is the heart of the matter. AMD is valued not much above its asset value at ~19 a share. AMD has a legitimate shot, albeit small, at taking a chunk of a market that yields Intel quarterly profits in the same range as AMD's entire market valuation. What will happen to that market value if the investment community begins to believe? What will happen if instead they continue to not believe? How does that risk/reward ratio look to you? To me it looks compelling enough that I have freed up a significant portion of my assets for possible use there. However, since Intel has inconveniently supplied me with a hard-to-judge benchmark, I am waiting until the approach of the 25th provides more information, because at that point I think we will have seen all the trump that Intel has to play until late next year. Intel is a fine company on which I have made good money. If AMD can become 1/10 as good a company, can it achieve 1/10 the market value? 1/20th? How much appreciation would that be?
Vince |
