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: Scumbria who wrote (7382)	9/2/2000 2:42:34 PM
From: Tony Viola	Read Replies (2) \| Respond to of 275872

Scumbria, >...but rather the tradeoffs (double speed ALU, 2-cycle cache) that potentially impact clock speed. I don't think you commented on the premise I made (apologize for not remembering the thread), and I saw Damone later say the same thing, that most instructions are very simple, like add and subtract, and can be done in one or two levels of logic. Hence, 2X on clock freq. may not slow down overall cycle time at all. WRT the 2 cycle cache access, do you know that it's a bottleneck, especially with a small cache? Are you privy to Intel's SRAM speeds? A lot of assumptions are getting made here, on P4, without enough data, IMO. Tony

To: Scumbria who wrote (7382)	9/2/2000 6:24:57 PM
From: Charles R	Respond to of 275872

Scumbria, <The big question for P4 will be clock speed. If they can ramp it up quickly to 2GHz, it will be competitive with AMD's offerings. > In my mind Willamatte MHz ramp in the short-term, even if possible (and I don't think it is), would not help Intel. A rapid race on the MHz side will orphan PIIIs and there is no way Intel can shift the production to P4 overnite given the massive die-size. Just for the sake of argument, imagine that P4 is at say 1.8G in Q1 and Mustang/Palamino is at 1.5G and PIIIs are yielding in the 0.7-1.1G range. What would Intel's mix of PIIIs and P4s be and what would the overall ASPs be? Intel's best interest is in NOT ramping the MHZ too fast until Willamette can be produced in high volumes (i.e. greater than 10Mu per quarter). What Intel really needs in the short term is an IPC delta in the 10-20% range. If it can be obtained with benchmark optimization, compiler optimization, additional caching, etc, Intel can do fine otherwise we are going to see some very interesting things happen in the marketplace. <Had they gone all out for GHz, (not worrying too much about IPC), they might have hit 2-3GHz right out of the chute, which would have made Willy a killer product. > Again, this would have been true if Intel could switch to Willamette overnite - a die size in excess of 200mm2 makes that highly unlikely even if Intel stops making a lot of other products (PIII chipset market has already been given up, will Intel giveup P4 chipset market too?) Chuck

To: Scumbria who wrote (7382)	9/2/2000 7:40:26 PM
From: andreas_wonisch	Read Replies (2) \| Respond to of 275872

Scumbria, Re: Had they gone all out for GHz, (not worrying too much about IPC), they might have hit 2-3GHz right out of the chute, which would have made Willy a killer product. I don't think so. If they would hypothetically release P4 at 2.0 GHz this October there would be a huge MHz delta between their fastest P3 and the P4. Of course nobody would want to have a 1.0 GHz P3 if a 2.0 GHz P4 was available. And because of the big die size Intel couldn't produce enough P4s so they'd actually kill demand. They can't move P4 to main stream until they are on 0.13 micron. And don't expect this to happen before end of 2001 at the earliest (I think Chic made a good point here concerning the difficulties of IBM and AMD with copper). Even then they will only have on (new) copper fab available for that process. Andreas