As gimmicks go, it's hardly as memorable as the fall Comdex a decade ago when one software developer decided to dress up some of its booth babes in the scantiest of tiger suits.
Still, it sufficed to remind the multitude of geeks and press types in attendance that there is an alternative view of the computing universe.
No secret why AMD's feeling spunky. For most of its decades-long competition with Intel, the company kept winding up with the short end of the stick. But now AMD is basking in its reputation as the technology trendsetter. Beating its bigger rival to the market with 64-bit technology, AMD also won over important segments of the tech cognoscenti on the strength of its Athlon and Opteron chip designs.
The company's also pushed Intel into a PR corner, accusing it of being a predatory monopolist. If the company's lawyers can make the charges in its lawsuit hold up in court, it's anybody's guess what the courts would do. (Had it not been for the reversal ordered by an appellate court, Microsoft, the other half of the Wintel duopoly, would have been busted up by Thomas Penfield Jackson, the judge overseeing the government's antitrust case against the software company.)
But so much for the guerilla theater. While AMD made the most of the current opportunity, Intel also made clear it expects next year will be a lot different. The company's general manager, Pat Gelsinger, went so far as to guarantee Intel would command "absolute leadership in the marketplace" in 2006.
Nobody's ever going to confuse Gelsinger's guarantee with Joe Namath's. But the company has its dander up and AMD's success has only sharpened the edge.
So it is that Intel next year plans a barrage of microprocessor releases that feature lower power consumption, more security against worms and better power management. Get ready for a huge advertising campaign where Intel plays up the energy savings and extra battery life.
Moving the debate away from pure clock speed is a clever marketing move, one that plays to Intel's advantages. It might also be an idea whose time as come. The numbers crunchers, the rocket scientists and the early adopters will still demand the latest unit off the production line. But most computer users already have enough processing speed to do their jobs.
My hunch is that more people will list items like low heat and power consumption higher as priorities. AMD needs to go all out in coming weeks to disabuse the notion that the speed race is over and the power race has begun--or else come up with a technology response before Intel ships its next generation of products.
If AMD fails that challenge, then Intel's latest gambit may wind up being recalled as one of its more brilliant moments.
Biography
Charles Cooper is CNET News.com's executive editor of commentary.
See more CNET content tagged:
Pat Gelsinger, AMD, Intel, AMD Opteron, clock speed





Technologically speaking, Intel is backpedaling all of their architectures to pre-P4 designs for Conroe, burying NetBurst and its current multi-core layout.
So, why, exactly, does AMD "need to go all out in the coming weeks"? They've got at least 12 months, the way I read the Intel roadmap.
Wow this report should check his facts before just writing this garbage!
Has Mr. Cooper been locked up for the last 4 years and has he been following developments in the Processor industry for very long? that is the only explanation for this article unless Mr Cooper is in fact on the intel payroll.
As stated by multiple previous posters, AMD has been touting lower clocked chips with higher performance since around ~2001.
I plead with you to please secure some writers that are at least somewhat in the know on topics they choose to write about instead of passing off this junk on readers that might not know any better.
In closing I must ask if Mr. Cooper has recieved his check from intel this month as this obviously is so biased and untrue that only someone paid by intel could have composed it.
so, the message to consumers between the lines is: "speed won't help you compute faster so we won't bother you with those silly numbers; just buy a new computer when the old one breaks down." that isn't exactly the smart way to grow a market. but, hey, it's good for consumers!
mark d.
mark d.
Having played with Longhorn for almost six months, I'm entirely confident that the next 3 years' of desktop computing power (and hard drive spaces) will be accounted-for.
-Matt
(overclocked to 2GHZ)
http://tinyurl.com/737fa
A few high lights
1) Look at CPU-Z, there is no 64 bit support in
Sossaman
2) 2 Sossaman32s (total 4 cores at 1.5GHZ) basically
tied with Athlon 64 4800+ which has only 2 cores
3) 2GHZ Yonah dual core with 2MB cache is 20% slower
than the entry level Athlon 64 X2 3800+
No wonder AMD calls for a duel....There is no way
Sossaman can compete with Opteron 280/880 at 2.6GHZ. On the power consumption front, AMD is selling dual core Opteron 865HE at 55 watts.
So if you look at the INTEL roadmap, its products of 2006 is far behind AMD products in Aug 2005. That is not a pretty sign.
Also, if you buy any INTEL P4 based products, they will be 100% obsolete in H2 2006, when INTEL moves to Pentium-M based architecture.
Like Dell "arrangement", Intel has blocked or dissuaded many of its OEM partners from using AMD products since the arrival of the Athlon -- AMD's first truly competitive product.
If AMD beats Intel at any given price-point and delivers more power-per-watt than any P4 chip, then why aren't their chips available in more systems?
AMD forced Intel's hand with DDR, with 32-bit-compatible, 64-bit processors, and now with multi-core architectures. When does AMD get a break?
1) AMD has designed x86_64 in 1999 and INTEL was denying the need for it. Now INTEL copied x86_64 from AMD on some but not all of their chips. INTEL's copy of x86_64 will be complete in 2H 2006.
2) Multi-core, AMD designed Opteron to be multicore in 1999, and INTEL only started in 2004. INTEL admitted its work was a kludge, see "Intel's dual-core Pentium 4 a rush job": http://www.macworld.com/news/2005/08/17/dualcore/index.php
3) Focus on performance not frequency. AMD has been doing this since 1999. INTEL moved from 12 pipe stages of P3 to 22 stages on P4 and 38 stages to Prescott, to bump up frequency (not performance), INTEL was talking about 20 GHZ in 2003 based on the Netburst. Now, INTEl moves back to P3 and 12 pipe stages and 2 GHZ.
4) Performance-per-watt. AMD and SUN brought up this concept and measure in 2004. See my previous post. INTEL has a 300% disadvantage to overcome.
5) Direct Connect Architecture. AMD has done away with the obsolete FSB design and use direct links between CPU and CPU, CPU and I/O, CPU and Memory. INTEL won't have this until 2008.
6) Embedded memory controller. AMD has this designed into every opteron and athlon64 and sempron64. INTEL needs a separate chip for memory controller. INTEL is rumored to copy this AMD idea in 2007.
- I agree...
- by Mendz August 28, 2005 5:49 PM PDT
- Software is mostly limited by its hardware. Hardware should be the primary source of power saving and security features. Having these features in place on the hardware level should be more effective rather than relying on software alone.
- Reply to this comment
-
(19 Comments)Though speed is still the name of the game as software development continues to progress, it is about time that hadware companies address the prevailing issues on power saving and security.
With the continuing sales growth of mobile devices, Intel wins for focusing on power saving features. And with the continuing threats on security, Intel wins for focusing on security features.