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: nuclear who wrote (229994)	4/15/2007 8:20:21 AM
From: j3pflynn	Respond to of 275872

nuclear - re: "The purpose of integrated graphics have been to enable low cost perform. intel current business model is for enable sales of CPU. Using your argument, AMD should have integrated a R600 class 128 shaders monster into 690G IGP instead of using a 4-pipe RV410 derivate. By the way, intel chipset use trailing edge process."<?i> If that's the case, why does AMD have the better-performing and less expensive IGP chipset as it is?

To: nuclear who wrote (229994)	4/15/2007 8:48:05 AM
From: pgerassi	Read Replies (2) \| Respond to of 275872

Dear Nuclear: Its an advantage because AMD will be using immersion and Intel will use double patterning. So AMD will have productivity advantages, ie they get more good dies per quarter per employee or per capital. Second AMD will get better definition and higher performance, lower power and/or leakage. You fail to note that while 65nm to 45nm will be about two years apart, 45nm to 32nm will be three years apart (three years of 45nm being cutting edge). While Intel would always be one node smaller in the past, they will have to suffer being at the same node as AMD for longer and longer periods. Last times it happened (for short periods), Intel was losing badly to AMD. Turn your argument around. AMD doesn't need to get to the smaller node first to beat Intel up. Going smaller isn't an automatic clock and/or performance boost it once was. Prescott showed that at 130->90nm and Cedar Mill for 90->65nm. The purpose of integrated graphics have been to enable low cost perform. intel current business model is for enable sales of CPU. Using your argument, AMD should have integrated a R600 class 128 shaders monster into 690G IGP instead of using a 4-pipe RV410 derivate. By the way, intel chipset use trailing edge process. But Intel's GPU on its most advanced process can't beat a integrated value older generation GPU on an older process. Something akin to the 610 will be in the 790G chipset and will run circles around Intel's fastest GPU while still not being on AMD's cutting edge process. TSMC's 65nm logic process is not cutting edge as AMD's 65nm SOI dual strain SiGe CPU process. Previously on this thread we had a long running discussion on what Fusion on AMD's 45nm process would be able to do relative to shrunk Barcelona to 45nm core sizes (single core with L1 being replaced with a x pipe R600 derivative with AVIVO). IIRC, 4 R600 pipes would easily fit into a core and clock four to five times higher. It would be the equivalent of a 16-20 pipe R600 which is far faster than any Intel GPU has shown to date even though its already on Intel's best 65nm process. It would be a mid range discrete GPU. AVIVO itself in its area would equal what a top end C2D can do using far less power. Besides as Intel well knows, it is far easier to remove functional units and reduce performance, than to multiply them to increase performance. Pete