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 (254802)	7/25/2008 2:32:42 AM
From: dougSF30	Read Replies (1) \| Respond to of 275872

It will be amusing to watch Pete's reaction to the Nehalem disclosures at IDF.

To: eracer who wrote (254802)	7/25/2008 4:44:18 AM
From: pgerassi	Read Replies (1) \| Respond to of 275872

Dear Eracer: Those benches were contrived by Anand. He is a big Intel backer as you can see the link there on nearly every page. No real single threaded tasks. No games. 10% more at the wall is larger than 10% more at the processor would have been, yet Anand seems to think that wall increase equals processor increase which isn't true at all. Doug seems to think that 12% less at the wall is less than that at the processor level. Usually any difference turns out to be much more at the processor level because of the power usage of the rest of the system sans CPU doesn't change. If you run the numbers considering PS efficiency and VRM efficiency with the only thing being changed is the CPU on the Shanghai side and Nehalem moving the MC and NB on die, that 10% more is like 35-50% more at the CPU for Nehalem. That would push the 95W TDP of a 2.66GHz Penryn based QC to 130W to 140W TDP for Nehalem. Which puts it in the same range of a Barcelona based 2.6GHz Phenom X4 9950BE. If you do the same for Shanghai, you get that it uses between 20-33% less than an equally clocked Barcelona. That puts it at 95W to 110W TDP at 2.6GHz. Quite a bit better than Nehalem that was broken according to Anand. Also note that Anand never said what load Nehalem was running when the power draw numbers were taken. I looked for that, but it wasn't in the "preview". That preview looks to be Intel propaganda far more than any real preview. 4 multithreaded tests using software that is known to be heavily Intel biased using some new switches (enhanced multithreading, meaning recoded for just this preview). One single threaded test of a streaming type benchmark. The other memory test was on a different faster CPU altogether. And that because the original tested CPU produced results that lost to the C2Q. An unbiased preview would have shown those results and stated it was a work in progress. Given that, there is a definite feeling these were selected because the results were quite favorable. Any unfavorable ones were bypassed. The Deneb benchmarks were done by a truly third party. They were standard ones done to all systems and CPUs therein. No playing of favorites was done. Both bad results and good ones were shown. Besides, AMD always seems tight lipped about unlaunched products even back to the early Opteron days. As to pricing, while there may be some cuts, I don't think we will see 3.2GHz C2Qs going for less than $286 after Nehalem (plus MBs) becomes available at retail long after its paper launch. Pete