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: Ali Chen who wrote (227050)	3/1/2007 3:25:06 AM
From: Petz	Respond to of 275872

Ali, it took spec.org six years to revise CPU2000 benchmark to make it more accurate. But no, you are absolutely sure that it is a bad benchmark. Your idea of comparing server chips using SPECint and SPECfp ragher than SPECint_rate and SPECfp_rate is beyond belief. In reality, SPECint_rate and SPECfp_rate correctly mimic the "real world" in showing how poorly Xeon's scale when you have more than 2 sockets. Do you think this is an accident, or do you think that maybe the SPECint_rate benchmark correctly stresses the poor connectivity of Intel Xeon server CPU's and chipsets? Petz

To: Ali Chen who wrote (227050)	3/1/2007 9:58:58 AM
From: pgerassi	Respond to of 275872

Dear Ali: Considering that AMD has flat out stated that Barcelona will be 42% faster in FP than Intel's fastest Quad core Xeon, I think you are the one who is dreaming. Or would you prefer the whistling past the graveyard image? aceshardware.com And this shows peak performance is better than you claimed, yet when SPECxxx_rate2006 is used, there is a dearth of C2D submissions. There are no submissions for multisocket C2D SPECfp_rate2006 at all. Given that servers do more than one thing at a time, especially as virtualization takes hold, this shows that C2D isn't really good for multitasking servers. Even the zero interthread communication versions of SPECxxx_rate2006. You showed the link. Turn on CPU MHZ and CPU count switches. Now show me multisocket C2D SPECfp_rate2006 scores. Whoops! Big hole in your "C2D is great" dreams. One multisocket C2D SPECint_rate2006 submission (Supermicro) and none for Intel compared to 6 for Opteron. And 5 for Opteron in SPECfp_rate2006 with one even beating a Itanium 4 core in FP compared to zero for C2D. Pete