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 (170003)	8/25/2002 1:18:53 PM
From: Dan3	Read Replies (2) \| Respond to of 186894

Re: I can't wait until you start denying SuperPI as a "valid benchmark" Why? You think you've made some kind of point, here? SuperPI is a program that was developed prior to the AMD vs. Intel architecture battles, it wasn't developed with either an Athlon or a P4, in hand, nor was that an issue at the time. There was no attempt to "optimize" for either architecture. It's a well known, robust, program that was written to be able to calculate very large values for PI. It exercises many parts of a CPU's architecture, and test results scale with clock speed, cache, FSB, and latency - it tests a number of aspects of CPU performance, and produces reasonable results as any of these aspects of CPU design change. It has all the characteristics of a very good benchmark. It's certainly not the only benchmark that should be used in evaluating processors, but it's an honest one, and is certainly a lot more representative of actual CPU performance than BAPCO's tests. I suppose you'd prefer BAPCO style crooked benchmarks that show a 100% increase on P4 performance with a 30% increase in clock speed, but produce poor results for non-Intel chips. Why do we never, ever see it in reviews? Do you think it has anything to do with the fact that SuperPI, almost all programs, uses X86 floating point capability, and P4's designs flaws have given it terrible X86 performance in floating point? How has Intel managed to make this honest benchmark vanish from reviews, while BAPCO is featured almost everywhere? It just happened? The most recent direct performance evaluation we've done was made to select systems for a team developing large energy sector models. The models use a combination of Java and Fortran routines and can take several hours to complete a run (depending on the scenario). We compared an Athlon 1700+ (1.47ghz) to a P4 1.8ghz system and found the Athlon to be 20% faster. That means that an Athlon at 1.176ghz offers the same performance as a P4 1.8 on a large, real-world application that include a variety of code. Athlon offers about 50% better performance, per mhz, than P4, for this large, complex, application. That's not too far off from what we see in SuperPI tests: www16.big.or.jp and a much, much, better indicator of performance than BAPCO.