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: jcholewa who wrote (5462)	8/17/2000 5:16:47 PM
From: Scumbria	Read Replies (1) \| Respond to of 275872

JC, There is of course the possibility that Intel took all this into consideration, and provided the logic to run the ALU at 1X speed. It may be as simple as toggling a configuration bit. But I doubt it ;^)) Any Intellabees want to break NDA and spill the beans? Scumbria

To: jcholewa who wrote (5462)	8/17/2000 5:32:13 PM
From: kapkan4u	Read Replies (1) \| Respond to of 275872

<> Besides the difficult datapath issues associated with the double speed ALU, the control logic is likely rather complex. > Removing the double speed ALU would probably be a 6-12 month effort. > Intel does full custom design, and any change like that is a major hassle. One of the big advantages that AMD has is they > use a cell based datapath methodology, which allows for much faster design turnaround times. Okay. I'll erase thoughts of dissolving FCLKs from my head, then. :)> I don't see where 6-12 months effort estimate is coming from. Double speed ALUs can also run single speed without any changes. P4 already has temperature control circuit that throttles the clock down when it is about to fry itself. The changes to the corresponding pipeline stages should be minor. The problem of removing double pumping is on the PR side. What would be the official chip frequency? They have three frequency domains now: 1. L2/decode half frequency 2. Trace/L1 fetch/execute full frequency 3. ALUs double frequency Today they can argue that the full frequency is a fair/average definition of the overall frequency, even though there are much more transistors in L2/decode than in the ALUs. If they remove double pumping then they will have to quote a 3/4 frequency instead of the full one. Kap