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 : Intel Corporation (INTC) -- Ignore unavailable to you. Want to Upgrade?

To: Tony Viola who wrote (109083)	8/31/2000 12:06:08 PM
From: Rob Young	Read Replies (1) \| Respond to of 186894

Tony, "Important to business?" Sure... you take every 200 million dollar sale you can get. Wouldn't it be better to have 1500 - 1 million dolla sales? Absolutely. But still you don't turn those down and a strong hint of things to come. Remember, we can get all sorts of defensive about what took place but Compaq wasn't a shoe-in. Better price and performance won the day. Case in point.. study the Pittsburgh deal.. for twice as much you can get the same IBM solution. "benchmark to compare the big players?" Well vapor to vapor? Sure... CPU to CPU bandwidth ;-) research.compaq.com While the above doesn't have the exact bandwidth numbers it is apparently higher than first proposed. I believe it may hover around 10 Gigabytes (no time to track that down) but why split hairs? The above paper is instrumental in proving a point as it introduces a new paradigm that others follow in the wake. Paradigm? Yes, Point to Point CPU transfers. Point to point CPU communication as on-chip L2 is very much extended in capabilities. Giving the very low latency, 2 hops or less away for neighboring L2 is very very effective.. consider what that means if you then see 16 CPUs sharing 1.5 MBytes * 16 or 24 MBytes of effective L2 per CPU. With RDRAM and on-chip memory controller whereby each CPU manages a single bank of memory (4 Gigs) you can see tremendous possibilities. Most of which is leaving out the VERY costly L3 and having extreme performance. "low" cost high performing as far as I can tell .. where "low" is a relative term ;-) (Actually, a 4 CPU box may be a very good sweet spot but we are a year out... futures!!!) But never mind me.... read the paper! Rob

To: Tony Viola who wrote (109083)	8/31/2000 1:00:13 PM
From: chic_hearne	Respond to of 186894

Re: This has me, a computer type, a bit confused. Would you care to set aside the tongue in cheek for a sec. and rank, IYHO, Itanium, Alpha, Power4, USIII, just for kicks? Pick some benchmark metric that you think is fair (if possible). We won't hold you to any of it, just for grins, you know? Tony, I'll pick java benchmarks, which is likely what you would be looking at for a webserver. 1). POWER4 2). Ultra Sparc III 3). Alpha 4). Itanium The Itanium is a guess. I've read technical documents by people that are just dumbfounded as to why Intel would go with an EPIC architecture. If they're right, Itanium will be an absolute dog at multi-threaded applications. chic