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: Tenchusatsu who wrote (161277)	3/6/2002 4:46:08 PM
From: wanna_bmw	Respond to of 186894

Tench, Re: "My point in all of this (and I do have one, believe it or not) is that the performance picture is very complicated. No one should believe that there can be one or two numbers that can accurately sum up system performance." And benchmarks have been evolving for the last 20 years, too.... I wonder what AMD thinks they can do with the TPI? QuantiSpeed is supposed to be the interim rating, right? AMD may need to hold on to that for the next 20 years, since as you say, developing the ideal benchmark is not an easy task. wbmw

To: Tenchusatsu who wrote (161277)	3/7/2002 3:06:25 AM
From: Joe NYC	Read Replies (1) \| Respond to of 186894

Tenchusatsu, my rating scheme would apply to the SYSTEM, not the processor. Much like PC Magazine compares actual systems (as opposed to processors like AnandTech or Tom's Hardware), the rating scheme would and should only be used to judge overall system performance. I think it is a good idea. I don't think this should preclude the hardware site from the testing game. There is no difference between Compaq or Del model XYZ and motherboard A, CPU B, Hard disk C etc. But the rating systems would have to be set up for components. For example, an online store which lets pick and chose components for a system would need to give you updated scores (of your benchmarks) as you add or change components. To do so, the online box-maker would need to have performance characteristics of all the components that he could plug into a formula, which would calculate the final rating (rather than having to test all the combinations of components that can make up a system.) The performance characteristics of the components could be lines, curves, arrays, or even simple scalar number. Thinking about this more, in case of a (CPU, Say Pentium Tualatin 1.2 GHz 512K L2), tfor a CPU, the performance number could be a single number. Well, I don't have this thought through, but the bottom line is that given a set of components with their performance characteristics, you should be able to plug the performance characteristics to a formula which would calculate how fast the system would run your 2 benchmarks. There are some interdependencies that would need to be taken care off. For example a CAS-2 DDR vs. CAS 2.5 DDR performance contribution would be different if they were in Duron system vs. Athlon, etc. Of course, my scheme is highly idealistic. Cheating would be rampant, and there would be no way to keep all the OEMs honest, especially the small screwdriver shops. True, but there is a way to prevent cheating. The benchmark would have to downloadable, so the first thing you would do is to download it and run it on a new system if it lives up to what you were sold. Joe