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.

Politics : Formerly About Advanced Micro Devices -- Ignore unavailable to you. Want to Upgrade?

To: Ali Chen who wrote (26283)	12/1/1997 1:10:00 AM
From: Maxwell	Read Replies (3) \| Respond to of 1572751

Enough of the argument here: The answer to Yousef-Ali's debate is that both of you are correct. The speed of the chip depends on architecture (pipelines and # of gates), layout (capacitance and resistance), and process (capacitance and resistance). Ali argues from the architecture side and Yousef argues from the process side. The bottom line is that all 3 factors interact with each other and often difficult to sort out which is the speed limiter. The only test to determine which process is better, AMD versus Intel, is to give each company a mask set of a device and let each of them manufacture. Since Intel is not using local interconnect the process to be used is 4 metal layer with no LI. Backend must be kept fixed and only front end (transistors) can be twiked such as reduction in gate. Standard TiSi must be used, not cobalt silicide or other better conductive material. Dielectric used is SiO2, no low-k stuffs. Metal must be aluminum. The ones that can push the speed the highest has a better process. It is silly to compare K6 and PII speed because they have different architecture and layout. Maxwell PS. Yousef: you have updated your SI profile. Very interesting. I've been to Purdue many time. Nice school, bad football team but an excellent basketball team.

To: Ali Chen who wrote (26283)	12/1/1997 1:28:00 AM
From: Yousef	Respond to of 1572751

Ali, Re: "We are clearly talking about two things that differ by two orders of magnitude in complexity and in importance." Oh yea of "little understanding", Ali ... one must first understand and optimize the basic building blocks before moving on to the complex system. It appears that you favor AMD's approach of designing the system first and then worrying about the process. Maybe this is why AMD has yield problems at the 233mhz bin point due to poor device design and process control. To put it in terms you might understand ... would you make the foundation of a 40 story skyscraper out of wood !!?? ... or would you use good engineering ??!! Please study more about the fundamentals first, Ali. Make It So, Yousef

To: Ali Chen who wrote (26283)	12/1/1997 3:34:00 PM
From: Petz	Read Replies (2) \| Respond to of 1572751

Ali, you are right about overall CPU architecture being as important, and maybe more so, than CMOS gate design. The differences between AMD and Intel performance are primarily the result of design tradeoffs, not superior/inferior technology. AMD opted for a lower latency and slower clock speed, Intel opted for higher latency and faster clock speed. AMD used a larger cache and top level architectural improvements to overcome the lower clock speed. AMD used several technologies to try and minimize die size, but at the expense of greater manufacturing difficulty. In retrospect, these decisions were wrong for the 0.35 process (yield problem), but who's to say that Intel will not have difficulty with local interconnect, etc., when they start trying to make the Deschutes. As for the 0.25 process, I am very surprised that the 1.9 volt Tillamook CPU runs two clock speed grades behind the Pentium II and only matches the performance of the MMX. AMD's 0.25 process at 1.9 volts is supposed to equal the clock speed of the Pentium II and be two speed grades ahead of the MMX. Petz