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:
The Speed Race
Numbers and Test Configurations
Synthetic CPU and Memory Benchmarks
WorldBench5
3dsmax
Lightwave [8], Cinebench 2003
CineBench 2003 cont.
OCR: Abbyy FineReader
Gaming Performance
Idle and BIOS Power
3D Rendering Power
DOOM3 Power
3Dmark'05 and Power
Prime95 and Final Thoughts

Give Us Some Feedback on this Review

 AMD Athlon64-FX57
The Cost of Speed
(Review by MS June 27, 2005)
AMD Athlon 64 3800+ (Venice)

Summary

Despite the fact that dual cores are the latest buzz, the speed race for MHz appears far from over. Amongst the reasons are certainly marketing considerations based on the obviousness of the current benchmarking structure used in performance ratings - in other words, if there is some headroom left, it needs to be shown. Or maybe it is just another exercise in feasibility.

There is no free lunch and higher frequencies are expensive not only with respect to yields but also to overall CPU power consumption. We were talking about diminishing returns earlier and the FX57 is no exception. Regardless, the latest top model in the FX series may easily qualify as the fastest single core desktop CPU on the market, or were our projections wrong?


Important developments never occur as isolated events, and if they do, they rather quickly fade into oblivion. In the PC business, the probably best example is the dual VESA VooDoo5 that very quickly became some sort of a cult object attracting die-hard followers and still, on the scale of revenue and overall success never made it past a “beloved oddball” status. Arguably, there were other factors involved, particularly the takeover of 3dfx by nVidia but at the same time, the VooDoo5 / 6 fable is somewhat symptomatic for a number of great technical developments that rose and fell within the blink of an eye – even on the admittedly short PC history time scale.

Despite rumors and white papers on dual core processors, dating back to 1989 with names as important as Pat Gelsinger on the list of authors, parallel processing in the desktop space has remained an oddball – until a few months ago. All of a sudden, however, we witnessed the event of nVidia’s SLI and ATI’s Crossfire on the graphics front and simultaneously the budding of dual core CPUs in both Intel and AMD’s product lines.

Dual cores are technically superior to single cores - just like dual cylinders on a car are better than a single cylinder, dual systems just run a tad smoother. The problem with smoothness is that it is easy to convince automotive customers. In the case of PCs where the entire marketeering strategy has focused solely on MHz or “Performance Initiative” ratings, it would feel somewhat odd seeing Hector Ruiz or Craig Barrett talking about smooth operation of their CPUs. Especially, if there are few benchmarks to back the claims – even though that particular aspect has seldom been an obstacle for marketing.

Bottomline is in the end that even if the days of single CPUs are somewhat counted, there is still a tremendous prestige value associated with them since there are benchmarks to prove the point and, who knows, at some point, the CPU may even go down in history as the fastest single core ever.

Another Star is Born?

On the other side of the coin is the issue of power consumption and power consumption is associated with thermal dissipation and electric bills, not to mention the environmental impact of PCs on global warming. Oh, and then there was the noise levels associated with cooling, especially in Media PCs and Home Theatres a huge factor.

It is exactly in that area where AMD has as of lately made a huge progress, the Venice core, albeit blemished with still a few errata (that hardly have an impact on the end user) runs at power levels that were hardly conceivably only a few months ago and is running cool enough to almost drop under the radar of fan audibility.

As we pointed out in our Venice review, though, the low power levels change quite dramatically as soon as the CPU is overclocked. In fact, the power consumption increases in a disproportional fashion so that a 17% increase in clock speed results in an almost 50% increase in power under load and possibly even more at idle (but keep in mind that at low power the systematic measuring error is higher).

It is no secret either that larger caches burn more power, for example, the power delta at equal clock speed and load between the ClawHammer and the Newcastle cores were approximately 10% more for the ClawHammer in raw Watts measured at the CPU. In so far, one of our points of focus is going to be on the power comparison between an overclocked Venice (at 2780 MHz) and the new FX57, sporting the San Diego core with the 1MB L2 cache and running at default speed – which, in this case, turned out to be 2814 MHz.

Athlon64 X2-4200+
(dual core)

next page: => The Test Configuration =>

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