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 : Advanced Micro Devices - Moderated (AMD) -- Ignore unavailable to you. Want to Upgrade?

To: wanna_bmw who wrote (57533)	10/6/2001 10:02:02 PM
From: combjelly	Read Replies (1) \| Respond to of 275872

"my only point in this, which I think has been lost, is to dispel the silly rumor that Athlon XP has model numbers based on how much faster it is than the Athlon (sans XP)." Based on a document on AMD's website, it is quite conceivable that AMD is using SPECint and clock-normalizing it. The Athlon XPs have about the right percentage better score than normal Athlons on this particular benchmark. Not saying this is a real valid way to do it, but...

To: wanna_bmw who wrote (57533)	10/6/2001 10:06:16 PM
From: kapkan4u	Read Replies (1) \| Respond to of 275872

<In terms of model numbers over megahertz, the Athlon XP 1500+ is rated 12.5% greater than the core frequency, while the Athlon XP 1800+ is rated 17.5% greater than the core frequency. Clearly, they are not being modeled after the difference between Palomino and Thunderbird cores> Check the SPECint2000 difference between Athlon and XP at the same frequency. The advantage is about 15% and grows with frequency due to hardware prefetch of XP. So 12.5% for 1500+ and 17.5% for 1800+ makes sense to me. Kap

To: wanna_bmw who wrote (57533)	10/6/2001 11:00:25 PM
From: jcholewa	Read Replies (1) \| Respond to of 275872

> a quick check of Athlon MP benchmarks with respect to a > similarly clocked Athlon show less than 5% performance > gains on many applications. > Athlon XP 1500+ is rated 12.5% greater than the core > frequency, while the Athlon XP 1800+ is rated 17.5% > greater than the core frequency. Clearly, they are not > being modeled after the difference between Palomino and > Thunderbird cores I believe that your argument here is not logically sound (even though your conclusion is realistic). On most applications, performance does not scale well in most applications. A 5% improvement is not completely unrealistic for a 12.5% boost in frequency. And it is always conceivable (though not proven) that the advantage of the Palomino over the Thunderbird will increase at higher frequencies.

To: wanna_bmw who wrote (57533)	10/7/2001 3:23:32 AM
From: TGPTNDR	Read Replies (1) \| Respond to of 275872

wanna, re: <Then again, I also agree with others on this board that think that arbitrary model numbers are a bad idea, simply because of this fact that they are arbitrary, and not based on any concrete standard of measurement.> This STUFF is all arbitrary(IMO). Think of how LITTLE meaning the CPU performance has in the scheme of things. By the time you shift MOBOS, HDDs, Memory, settings, Video, etc., then pick your test software, the tester can move the performance comparison results to a large degree. And what should be the measure of performance, anyway? The more distinct the test the farther away, in general, you get from 'real world', whatever that is. In my world it's large databases, GUI development in GUIs, and web apps. Others live in gameland. Some seem to want to run their screen saver as fast as they can[Why else would you optimize SETI?]. Whatever! In how much of what you do could you tell the difference, in accomplishment, over the course of an hour or day, between a PIII-500 and a PIII-1G? Can you notice it in web browsing? Word processing? I can certainly tell the difference when I'm compiling or running a GUI software development environment, but what % of the market does that KIND of stuff? When I'm looking at database & web servers I can see it, but a lot can be said in the parallel cheap hardware vs single box fast hardware argument. And the Athlon/PIV generation is almost over. tgpntdr