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: revision1 who wrote (35759)	4/14/2001 3:13:01 PM
From: ptanner	Respond to of 275872

Revision1, Re: Thermally limited P4 I guess the P4 would not be the best choice of a processor for a distributed computing project? Such as the Intel sponsored cancer cure project? Interesting since they use a P4 system as the basis for evaluating the "power" of participant's computers. My old P200 was a 17 (on 1-100 scale) while my Duron 600@945 ranges from 75-80 depending on the network value. -PT

To: revision1 who wrote (35759)	4/14/2001 8:24:33 PM
From: ted burton	Read Replies (4) \| Respond to of 275872

I am the only one on this thread with actual data. Everyone else is simply making assumptions (some better than others). The max power app was programmed and refined over a period of years by the architects most intimately familiar with the part's internal workings. It's only purpose in life is to ring the power bell on this particular architecture. No other app we've seen comes anywhere close to it. The apps you'd think of as power hungry (Ispec, Fspec, mpeg encoding, circuit simulation, gzip etc.) are actually very tame in comparison. They may tie up the CPU for hours on end, but they don't do what the power app does. The thermal management feature allows OEM's to design their thermal solutions around real apps without worrying about someone hitting them with an artificial power generating program. Designing this way would normally have carried the risk of blue-screening, but this feature changes the risk into a very rare, and very minor slowdown. This feature offers a huge advantage, which is really only vulnerable to FUD. Go ahead and hype it up - convince yourselves it's a major flub-up. Don't overdo it though because its a sure bet the OEM's have already asked your beloved AMD to copy it, and they've already got someone working on it. I rarely visit this thread & likely won't be back for months. Sorry I don't have the time to address all the comments individually. -Ted Burton Again, I am an Intel employee, but do not speak for my employer.

To: revision1 who wrote (35759)	4/15/2001 3:56:27 AM
From: Bilow	Read Replies (2) \| Respond to of 275872

Hi revision1; Are you routing Virtex stuff? Ali Chen swore to me that Pentiums were more efficient than Athlons on them. Also, I don't seem to be needing more than around 384MB for a design that 60% fills a V1000E. I don't see how you're realistically using up 2GB. My experience comparing different machines routing is that the most important thing is processor speed. And you only need to have just enough memory that you don't thrash the hard disk. But things change, and maybe you're using Altera... -- Carl P.S. I agree that the P4 having to throttle itself when it actually gets enough data to avoid stalling is a hoot. Where have we seen this before... Oh yeah, it's the same technique that prevents RDRAM from melting when a single chip is repeatedly read from, LOL!!! By the way, the most telling part of the McComas article was when he (more or less) compared the total number of bits moved over the memory interface by a Pentium III vs. a P4, when performing the same benchmark. That one statistic, that the P4 uses 4x the memory bandwidth to deliver 1x the performance, tells it all. When a single processor burns up all your memory bandwidth you can't expect much of a performance improvement from a dual. Which of course explains why the P4 isn't available as a dual. And you know that this sort of problem isn't going to be easy to fix. Intel is stuck with the P3 for dualies. Result: *AMD rulez!!!*