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: jcholewa who wrote (67519)	1/10/2002 1:18:25 PM
From: Ali Chen	Read Replies (1) \| Respond to of 275872

JC, "SuperPi's performance characteristics do not, in my opinion, model any real world scenario. They are almost totally pointless, especially if you're trying to analyze the effect of chipsets on system performance." I think you totally missed the point. In my understanding, Dan's message contained two important points: 1. Athlon platforms are capable to run at 210MHz FSB, supporting the idea of AMD being capable to increase the FSB in order to gain/maintain platform performance. 2. Lower-speed CPU with higher-speed FSB outperforms higher speed CPU, which indicates that the particular application has a fair amount of off-chip traffic, a feature that is not uncommon in many SPEC apps. The point was to illustrate that faster FSB could be more important than the core speed, and will be increasingly important given the growing gap between slow memory performance and blazing CPU speeds. The fact that the platform was able to run any sort of operating system is itself a big achievement and proof of concept. - Ali

To: jcholewa who wrote (67519)	1/10/2002 1:26:28 PM
From: Win Smith	Respond to of 275872

But running time of 48 sec. vs. 49 sec. in two extreme overclocked configurations, probably with LiqN2 or something, would model some "real world scenario" if only the proper benchmark was used?

To: jcholewa who wrote (67519)	1/10/2002 2:09:17 PM
From: Dan3	Read Replies (2) \| Respond to of 275872

Re: SuperPi's performance characteristics do not, in my opinion, model any real world scenario As opposed to what, windows media encoder? First of all, I know of no other set of benchmark results that can be used to compare FSB vs. clock speed performance. But, regardless, other than the fact that it makes P4 look bad, why is SuperPI not a good benchmark? It is a heck of a lot more "real world" than the legion of memory bandwidth numbers we see in every chip comparison. Calculating Pi is a task that represents a class of operations. There are no restrictions on how to code a solution to the task, and it is open source. It's a perfect benchmark for one class of CPU operation. Yet this benchmark is never seen in typical reviews because Intel doesn't like it. Windows media encoder, OTOH, is a specific, closed source binary that that performs a task that is also performed by other, similar programs. But WME gives strangely different results from other media encoding programs used to compare the same hardware. Still, we see the results of the anomalous WME in every review - it even is run in the background in the most often used benchmark suite, affecting the results of every benchmark in the suite. I think your conclusion is way off base.