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: kapkan4u who wrote (106214)	4/17/2000 1:43:00 AM
From: Scumbria	Read Replies (2) \| Respond to of 1573994

Kap, In effect, Willamette has an equivalent of a tiny ~16k L1 instruction cache, albeit a fast one. Programs with poor locality will run slow, while Willamette's single decoder is straining to fill the trace cache after trace cache faults. Sorry to rain on the anti-Willy parade but- PIII has a 16K instruction cache, and the performance is just fine. The penalty of a trace cache miss will be no larger on Willy than an L1 miss on PIII. ALUs transistors running at twice the clock speed generate twice as much heat as well. The ALU runs at the same clock speed as the rest of the chip. It just completes arithmetic operations in one phase and forwards it to the next phase. Willy shouldn't draw significantly more power than Athlon at the same speed in a similar technology. Willy will be a good performer, and run at a high clock speed. Let's not kid ourselves. Scumbria

To: kapkan4u who wrote (106214)	4/17/2000 2:49:00 AM
From: Charles R	Read Replies (1) \| Respond to of 1573994

Kap, <The uops density in the trace cache is 5 times worse than the corresponding x86 instructions. In effect, Willamette has an equivalent of a tiny ~16k L1 instruction cache, albeit a fast one.> Yeah! It looks highly likely that Thunderbird/Mustang will have better CPI than Wilamette for this and other reasons. <ALUs transistors running at twice the clock speed generate twice as much heat as well. The power has to be delivered and dissipated through the surrounding silicon. This sounds like a severe MHz scalability hurdle.> Looks like Athlon is at least as handicapped in this regard. The number of execution pipes in Wilamette are fewer than in Athlon. <If Dell goes with T-bird and Spitfire, then Intel stockholders watch-out. > If Dell goes AMD, AMD folks should forget about Intel and just party. Chuck

To: kapkan4u who wrote (106214)	4/17/2000 5:44:00 AM
From: Daniel Schuh	Respond to of 1573994

Kap, I'd agree with Scumbria for the most part on that one. My take on the Paul Demone article was that it was mostly favorable toward Willy, the 90k/16k analysis on the trace cache was just an estimate on where the 90k sizing came from. As Scumbia posted, 16k has worked good for P6. To get an idea about Willamette's performance watch for two things. First, if performance was good, SPEC numbers would have been leaked already. I agree somewhat with that point, though. "Intel does not comment on unannounced products", "except when they're running scared from AMD" seems to be a required addendum. I smell specially cooked benchmarks in the pipeline ala Rambus. Second, watch Dell. They have the earliest access to Willamette samples. If Dell goes with T-bird and Spitfire, then Intel stockholders watch-out. The bottom will be very hard to find. On the other hand, I think expecting anything good news from Dell wrt AMD is wishful thinking. An analysis that the Register made is that Dell is almost a distribution arm of Intel. They're not like Compaq, IBM, or HP, each of which has a lot engineering depth in building systems. Dell gets most of their engineering straight from Intel. Plus, Mikey's loyalty has put them at the head of the queue for all the scarce Intel parts, which has to be a valuable asset when the corporate world still want Intel inside. IF good news ever comes from Dell, I expect it to be later rather than sooner, and I wouldn't hold my breath at any rate. Cheers, Dan.