Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Dan3 who wrote (163138)	3/29/2002 12:45:24 PM
From: wanna_bmw	Read Replies (1) \| Respond to of 186894

Dan, Re: "Never mind what issue, P4's flacid, inadequate mhz?" Actually, AMD is the one with flaccid, inadequate megahertz. The Pentium 4 has flaccid, inadequate IPC. But in the end, both of these tend to balance out, giving the user a similar performance experience. So I don't understand what your beef with the P4 is. I think the issue here is whether QuantiSpeed gives AMD a better number to compare more favorably to the Pentium 4. IMO, QuantiSpeed is the half assed solution, since it continues to attach its baseline to the Pentium 4 measurement that is inadequate to begin with. Since megahertz is not a good measurement for performance (because it is lacking a very important ingredient), then why does AMD follow it with model numbers that are meant to match it? It only gives due credit to the one measurement that should be de-emphasized. People see QuantiSpeed, and see that it compares to megahertz, and that only reinforces their notion that megahertz is the right way to measure performance. This is not a good solution. Besides that, I have already outlined what I feel are the largest mistakes that AMD has already made with QuantiSpeed - and there are several. This diminishes whatever credibility they would have on establishing a comparative measurement. People are certainly going to question how AMD can simply "make up" model numbers to suit their current environment. How do we make sure that AMD is "fair" about it? Their current formula, ((frequency-333MHz) * 3/2), has no scientific basis. It easily showed a performance advantage over Willamette, but now it has greater trouble showing a benefit over Northwood. That means it will have to be changed. But to what? Is ((frequency-366MHz) * 3/2) any better? How does one "quantify" where the measurement should be set? Where does Arthur Anderson come in, since they only verify AMD's benchmark results - not the formula itself? In every sense of the word, AMD's formula is arbitrary, and that has serious consequences as things change over time. People will eventually have to question where the benchmarks fit in. It won't be good enough just to see AMD slightly ahead of the comparable Pentium 4 megahertz. People are going to want to verify that AMD set the model number as accurately as possible, but what is AMD going to do about that? They have no set procedure for setting the formula that they have disclosed. So far, it's been [ if (QuantiSpeed_Performance > Pentium_4_Megahertz_Performance) then {QuantiSpeed = ((Frequency-333MHz) * 3/2);} else {ThinkofSomethingElse();} ]. AMD will be even more scrutinized when the else statement comes into effect - which should be shortly in the future at their current rate. wbmw