Navigate:

Advice
Beginners
BIOS Guide
CPUs
Links
Mainboards
Memory
Network
Storage
Video/Sound Cards

Contact
Forum
SiteMap
Sponsors
WebNews
Home
. .

Prices:

Mainboards

ABIT
ASUS
Chaintech
Shuttle
Soyo
Tyan

CPU
Intel
P4 2.4C-800
P4 2.6C-800
P4 2.8C-800
P4 3.0-800
P4 3.2-800

AMD
AthlonXP
XP 1700+
XP 2000+
XP 2400+
XP 2500+
XP 2700+
XP 3000+
XP 3200+

Athlon64
Athlon64 3200+
Athlon64 FX-51

Opteron
Opteron 240
Opteron 242
Opteron 244
Opteron 246

Memory

Corsair
Crucial
Kingston
Mushkin
OCZ

Search Prices:


























































































































LOSTCIRCUITS

SHORTCUTS:
Top
Technical Details
Benchmark Considerations
Gigabyte GA-7DX vs. VIA Reference Board
Shuttle AK31 Rev. 3.1
Driver Issues, Test Configurations
Quake3 Arena, Expendable Timedemo
Caligari TrueSpace 4.2 Animated Benchmark
Caligari 5.1 Raytracing, Ulead MediaStudio Pro 6.0
Conclusion
 Athlon XP1800+   
QuantiSpeed in the Crossfire
(Review by MS, October 14, 2001)

Summary

AMD has released the desktop version of the Athlon 4 and Athlon MP with the suffix XP and a new performance rating scheme. The major technical difference between the Athlon XP and its predecessors is the new Organic Pin Grid Array (OPGA) to reduce manufacturing cost and open the way to new speed grades by decreasing I/O impedance. Other features of the Palomino core such as hardware prefetch and increased TLBs were already introduced with the Athlon 4 and MP series. New however is the marketing strategy, assigning an arbitrary Instructions Per Clockcycle (IPC) factor as more or less constant multiplier to the physical clock speed of the new CPUs. The resulting QuantiSpeed Architecture rating opens a wide door for criticism, justified or not, even though our benchmark results indicate that the numbers are rather conservative. Another issue is that with the new rating scheme AMD appears to acknowledge Intel's Pentium4 as the measure of all things in CPU circles.


The Athlon XP 1800+ side by side with my original Thunderbird 1GHz (the latter after a thorough bath in Pro-HandCleaner from PepBoys to remove ArcticSilver from several dozens of mounting in different systems). Note the different packaging using either fiberglass or ceramic and the rearranged core featuring some additional 500K transistors for increased TLBs and prefetch registers..

.

A few weeks ago some cries of agony were heard from a variety of different sources. Those cries were caused by the leaked news that AMD was going to resuscitate the infamous performance rating system that still causes some bad taste on some tongues. Upon getting in touch with a few of my contacts at AMD, the rumors were confirmed and justified as the only way to counter the GHz hype on which the Pentium4 soars through the public perception of the uneducated masses.

The counter strike constitutes the following eXtreme Procedures. First of all, a new name, that is AthlonXP. The meaning of XP has changed a few times since its original conception. My own favorite is actually not officially condoned by AMD but stands for eXtragalactic Performance to bring the computing power of the entire universe and beyond to your desktop. And then there was the coincidence of MicroSoft releasing their highly acclaimed new operating system with the same suffix (even though the meaning is somewhat different).

I believe it was Isaac Asimov who stated that Humans can only understand what they have already known earlier and the concept of piggy-back rides appears to be quite successful as we can tell from the number of (Name)Tech websites, no offense.

Sarcasm aside, here is the real situation: AMD finds itself in the situation that despite having a processor with superior performance, most of the marketing still concentrates on raw Mega or GigaHertz. As a consequence, Intel still stands as the performance leader, at least in the warehouses and with some of the OEMs. From an architectural and design standpoint, raw clock speed is relatively easy to achieve, add a few extra long pipelines is one possibility to increase the ease of internal data flow and get the clock speed up. This concept has been applied very successfully again in Intel's Pentium4. Performance is not necessarily a side effect of the elongated pipelines, though. Therefore and based on benchmarks, AMD claims a much higher efficiency of their CPUs as measured in instructions carried out per clock cycle (IPC). The actual workload performance is then calculated as:

Most people are really bad with formulas and without a real label assigned to the idea it would not stick. The new term coined to describe Application Performance is therefore:

An Invitation for Criticism

In other words, the arbitrary constant of IPC acts is used to adjust the physical clock rate to a performance rating. There are, of course, problems with this scheme and the biggest problem is that it can be criticized, regardless of whether it is adequate or not. I personally will not go into the pro's and con's of the discussion here since they are highly subjective and driven by marketing as well as stock holders interests rather than based on real considerations.

My only comment in this respect is that AMD, in their striving for the most powerful processor has missed its chance to cheat at the very first introduction of the Athlon. It would have been easy to stick in an extra resistor to filter out every other external clock signal and run at twice the multiplier to achieve a nominal 2x clock speed with only half performance. Since the available CPUs provide a standard against which every new revision can be measured, this is no longer possible without getting caught immediately. Therefore, my personal take is that the rating system is an honest defense in a frequency dominated marketplace, which, however, can potentially backfire quite badly.

next page:    => On to the Testing =>

Click here! All advice and educational articles on LostCircuits are free, but if you feel you can, please make a small donation to us!
Thank you!

General disclaimer: This page only reflects the author's personal opinion and assumes no responsibility whatsoever regarding any of the contents or any damages that may occur explicitly or implicitly from reading the contents of this site. All names and trademarks mentioned in this review are the exclusive property of the respective parent companies.
All contents of this site are protected by international copyright laws. Reproduction of the contents even in parts is not allowed except after written permission by the author and referral to this site.
Copyright 2002 - 2008 LostCircuits