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: pgerassi who wrote (202412)	6/15/2006 10:04:46 PM
From: Joe NYC	Respond to of 275872

pete, Sorry it does matter. Sun's best Opteron time in SPECfp2000 is with 2 cores enabled because they have and auto parallization option. What this does is use a 2nd core to run some independent FP calculations and/or doing one pass with the first core and the second pass with the second core. This seems to increase scores overall by 20-30%. It also helps that the second and higher cores can remove all OS load from the first core. I am aware of autoparallelization in Sun scores, but I am not sure that is what Intel is doing in SpecInt. OS load is negligible. What gives Woodcrest advantage in single thread apps such as SpecInt vs. multithread is large shared cache. All of the NGA processors will show this advantage in single thread apps. I pointed this many times in the past, that the shared cache will give Intel additional advantage on single thread apps, and now we see it. On the server (multithread) it will diminish somewhat. Socket F (1207) will double Opteron's available bandwidth per socket. There are good indications that this helps with all four benchmarks (int, int_rate, fp and fp_rate). Higher clocks are likely as well. Agreed, but the gap is fairly wide on SpecInt side. SpecFP will improve more than SpecInt. Joe