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: eracer who wrote (196949)	5/15/2006 10:28:17 PM
From: pgerassi	Read Replies (1) \| Respond to of 275872

Dear Eracer: I wouldn't call having only 60% of the Sciencemark 2.0 score being nearly as fast. In many of those benchmarks, Core Duo T2600 was behind the A64 X2 4800+ as much as an Opteron 165 or even a downclocked (1.8GHz) A64 X2 3800+ which in AM2 form is to be a 35W TDP CPU at 2GHz. Another point is that the S&M burn package might not use the most on Yonahs as it does for P4s as burn programs need to be specific to the target CPU. Each CPU has a different set of resources and if you use the wrong instruction mix, a bottleneck in one area will reduce usage in the rest. And the performance wasn't measured while the load was being measured. Thermal slowdowns are easy to detect when performance varies over time (usually down). Pete

To: eracer who wrote (196949)	5/16/2006 3:22:15 AM
From: Petz	Respond to of 275872

If you take those benchmarks and calculate the geometric mean performance, you get 4.239 for the T2600 and 4.610 for the 2.4 GHz A64 X2. That is ignoring the 3 synthetic benchmarks, Sciencemark, Mathematica and the TrackMania game, which clearly was at a video-card limited resolution. I also weighted the two 3dsmax scores by .5 each, since they are mostly measuring the same software. Speaking of 3dsmax, it seems peculiar to me that of 6 software titles used in the industry standard SPECAPC, they chose 2 of them and repeated 3dsmax twice. Why no PRO Engineer, SolidEdge, SolidWorks, etc? Sciencemark and Mathematica were thrown out since they were the most pro-AMD and pro-Intel of all the benchmarks, respectively. So with these unweightings, the performance difference is 8.75%. If performance scaled perfectly with clock speed, a 2.2 GHz A64 would have equaled the 2.16 GHz Yonah in this 32-bit code. However, we know that many of these bencmarks do not scale even 50%. In particular the 4 games, 2 Sysmarks (another Intel favorite), Premiere. ITunes may be memory bandwidth sensitive. Add a few % for the AM2 platform, and a 2 GHz Turion X2 is the equal of this top-of-the-line Yonah chip. And, guess what, probably lower power and 64-bit besides. Petz