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: KeithDust2000 who wrote (169667)	8/23/2005 2:25:01 PM
From: Rink	Read Replies (1) \| Respond to of 275872

Keith and all, the relative performance figures say NOTHING because: 1. Are they comparing max integer performance for mobile/desktop (and max TCP-C for servers)? 2. Or average performance at a range of wattages? 3. Highest performance per watt even if that is at a power consumption of an order of magnitude below TDP? 3. Native 64b performance versus indirect 64b performance using a 32b ALU? 4. Theoretical unknown Integer measurement tool specifically optimized for Merom/Conroe? 5. Etc... They are impressive numbers, but how impressive will they really be? Regards, Rink

To: KeithDust2000 who wrote (169667)	8/23/2005 2:50:51 PM
From: NicoV	Read Replies (2) \| Respond to of 275872

The graphs are titled performance/Watt. So what I understand from the graphs (taking desktop as example) is that the fastest speed grade of Conroe will deliver 5 times the performance/Watt of the fastest speed grade of Northwood (what was that? 3.0 GHz?). Their own graph confirms this, as Smithfield runs at about the same clock speeds as Northwood (let's say 3 GHz), but with significantly higher TDP (let's say 100 Watt versus 70 Watt). Assuming that Smithfield has perfect scaling (in number of cores), performance per watt would be 2 * 70/100 = 1.4, which is about the number of Smithfield on this graph. For the hypothetical example of comparing DC Opteron with Smithfield, just start with the observation expressed in many reviews that the fastest Intel dual core is slower than the slowest AMD dual core. Let's say a top speed Smithfield is about a 1.8 Ghz DC Opteron. Compare the Smithfield at 100W with a 55W DC Opteron 265 HE (the low power version): 100/55 =1.82. With these assumptions, a 265 HE would get 1.4x1.82= 2.5. Conroe will have a two times higher performance per Watt number, i.e. Conroe will be twice as fast as a dual core 1.8 GHz Opteron. This may be not so far (20%) from the fastest dual core of AMD in 2H 2006.