Are you eXPerienced?

Author : Wayne Date : 14th October 2001

3DVelocity would like to thank AMD and especially Theresa Zimmer for their help and courtesy in providing this processor for review.

 

 

Part One: The wait is over...

It seems like an age already since the speculation began about the new version of the Athlon chip. The so-called 'Palomino' has been on the tip of everyone's tongue for many months now. Rumor and speculation was fueled even more by the Internet's essential dynamic; freedom of speech. It wasn't long before the company's revamped naming procedure was released for public viewing, and while I would suggest it was known about a tad too soon for AMD's liking, it certainly did them no harm at all on the publicity front. I'll explain the naming later in this review, for now all we need to know is the 1800+ equates to 1.53Ghz part. AMD gave us glimpses into the future by building anticipation with the releases of the 'MP' Athlon and the laptop version of the uprated core, the 'Athlon 4'. The 'MP' (Multiprocessing') Athlon achieved rave reviews upon it's release proving to be the fastest multiprocessing solution available despite the lack of mature motherboard availability. The 9th of October turned out to be the release date for the XP as some have known for a fair amount of time now, so, as the section heading suggests...the wait is over!
The scene is set, the chip is here and what would you know? The XP found its way to 3DVelocity.com thus starting a series of exclusive reviews.

AMD Athlon XP; overhyped and underspecified or another winner for the AMD processing brand? Let's find out...

 

Part Two: A new standard?

Where to begin? I would guess that most people really need to know what is new in the XP Athlon before they are going to shell out money on another upgrade. I'll try to explain the main architectural changes without baffling you with needless company jargon and hyperbole.

To explain the changes we need to understand AMD's updated agenda. It's one that will no doubt cause countless numbers of disagreement within online communities and printed media alike. AMD's primary aim is to try to devalue the MHz factor and concentrate on other factors within the design of the CPU as a reference for performance.
One thing AMD and Intel alike have to agree on is the underlying principles of CPU performance. Simplistically speaking, performance of any central microprocessor is measured by an extremely simple equation.

(IPC (work per clock cycle) x MHz (clock speed)= A measurement of 'relative performance' )

As both Intel and AMD have started to vary the IPC AMD are, rightly so, stressing the importance of this in contrast to simple frequency increases. Indeed it's true that Intel chose, strangely, to break an important convention with the release of the first Pentium Four CPU's. The actual IPC went down not up as had been the case before. Every new generation of Intel CPU prior to the P4 had improved the IPC yet the P4's IPC went down by almost 20%. But Intel's main weapon lies in the naivety of the high street buyer. It's easy to observe the appeal a 2GHz chip yields but does it mean higher overall performance? That, if anything, would be AMD's main message to consumers. Indeed, it seems they have a great point: The places where performance counts the most, like servers and powerful workstations, still utilize relatively lower frequency CPU's. (The IBM 'RS/6000' (450MHz) would be a good example for this) With game specifications still barely touching 650MHz has Intel placed too much faith in pure MHZ? AMD would certainly like you to think so.
In the future consumers may need a different standard to compare CPU's but what system can be relied upon? AMD wishes to utilize a standard benchmark against which all CPU's will be compared, the updated naming policy is the first step to achieving this standard. After being, somewhat embarrassingly beaten to the coverted 1GHz mark, Intel are unlikely to downplay the frequency gap they have created of late. The XP versions of the Athlon will have to be strong performers to compete, are these chips up to the challenge laid before them?

 

Page Two: Major Changes?
>>

 

 

Home