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: NicoV who wrote (169690)	8/23/2005 3:09:19 PM
From: KeithDust2000	Read Replies (3) \| Respond to of 275872

NicoV, thanks for sharing the values you used. Now if you look at the comparison of Smithfield (130W) vs. Conroe (65W), then this suggests a huge performance increase, absolutely phenomenal, that AMD can´t even dream of approaching with A64 X2 even assuming ridiculously optimistic frequency potential. What am I missing?

To: NicoV who wrote (169690)	8/23/2005 4:10:24 PM
From: pgerassi	Respond to of 275872

Dear NicoV: A few notes about your calculations: 1) I do not think they used the fastest version for each generation but, the mean CPU speed bin. We all know that Prescott delivered less performance for more power than Northwoods shipping at the same time. But, the PPW (Performance Per Watt) number on the graph is higher for Prescott than Northwood. So either they are using a program that does poorly on Northwood or the Northwood base used is one of the earlier models that used the same power but, at lower clock. So the likely candidate is the 2.4B GHz Northwood in 2003 at 94W TDP (58W typ). 2) The Prescott used was likely the 3.0D GHz with HT at 101W TDP (84W typ). The HT, larger L2 and higher FSB probably helps enough to get a higher PPW number even though the power is higher and the IPC drops. 3) The Smithfield used is likely the 3.2GHz HT EE at 202W TDP (168W typ) version. That is why its gets a slightly higher PPW. 4) The slowest dual core A64 X2 (3800+) is much faster (15-25%)than the 3.2EE Smithfield on integer code (the basis in that graph). Its TDP is 95W (57W typ). This is where you get the biggest gap between your calculations and the real world performance. The new calculations should show 2.4B NW gets 2.4G/94W or 25.4MPW. The 3.0D HT P4 gets 2.8G/101W or 27.7MPW. The 3.2EE DC P4 gets 3.0G2/202W or 29.7MPW. Conroe thus should get 125MPW or the equivalent of 4G2/65W or somewhat lower than A64 X2 4800+'s performance at same power. The X2 uses about 70W on K8Burn. So A64 X2 4800+ already has the PPW of the future Conroe, if AMD rated it like Intel does. It already exceeds Conroe in absolute performance. For Intel to get back in the game, AMD would have to stand still for over a year and gain nothing going to 65nm, faster memory and possible asymmetric cores and/or additional FPU units. Not bloody likely. Pete