SI
SI
discoversearch

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: Sarmad Y. Hermiz who wrote (217983)11/29/2006 11:24:16 PM
From: dougSF30Read Replies (1) | Respond to of 275872
 
AMD is going to attempt to improve their 65nm transistors in 2007, as the quarters roll by. They need to get SiGe strain working, for one thing. Will they be successful? Time will tell.



To: Sarmad Y. Hermiz who wrote (217983)11/30/2006 3:37:17 AM
From: pgerassiRespond to of 275872
 
Dear Sarmad:

Don't read much if anything on the released clock speeds. 3 of those bins were vacated by 1MB L2 90nm components. AMD seems to feel that 100MHz of higher frequency will do better than double the L2 on the average. In this, I think they are right, slightly faster (4-5%) does beat (performance wise) nearly twice the cache, especially when considering die size, fab capacity and margins. The lower power will bring bigger benefits.

Given the manufacturing capacity constraints, it is better to replace the sweet spot CPUs where the resulting profits and market share are much better. After all, isn't it good to still make useful 90nm parts prior to Fab 30 to 38 conversion? When K8L appears, look for a quick down ramp of 90nm in the high end. Also note that the sweet spot Conroes are also at the 1.8GHz to 2.3GHz range, well below that of initial 65nm K8Gs (2.1-2.6GHz). That even allowing Intel a one year lead, AMD's average is already is beyond that of Intel's latest average.

My gut feeling is that in Q1, a lot of SKUs will quietly convert to 65nm with servers being the longest on 90nm given their reluctance to change.

Pete