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:








































































































































What are you
shopping for?



































































































































































LOSTCIRCUITS

SHORTCUTS:
A Darwin Award
Prelude to a Bug
Virtualization and TLBs
A Band Aid and a Patch
Phenom Specs
The Spider Platform
Test Configurations
ASUS M3A32-MVP and AOD
Memory subsystem
CPU Power Consumption
TrueSpace and Power Efficiency
Cinebench
DVD-Shrink, MainConcept
VirtualDub/DivX
3DMark'06
FarCry
F.E.A.R.
World In Conflict
Crysis
UnrealTournament3
NB Frequency: Does it Matter?
The Secret of AOD
Final Analysis

Give Us Some Feedback on this Review

 AMD's Phenom Processor - Beyond Erratum 298
(Author: Michael Schuette, January 1, 2008)

NorthBridge Frequency and System Performance

A major difference between the Phenom 9700/9900 and the lesser speed grades is that in the latter offerings, the Northbridge, including memory controllers and L3 cache, is running at 10% lower speed, that is 1.8GHz instead of 2 GHz. As mentioned earlier, the NB is running at fixed frequency and also has its own dedicated power supply controlled by the fifth phase of the hybrid voltage regulator. The whole idea is somewhat intriguing, assuming the memory controller runs at a more constant load than a CPU, one wouldn't need as elaborate voltage regulation. Moreover, it becomes possible to uncouple the NB voltage from the CPU supply voltage and, even if the core voltage is increased for overclocking purposes, it would not affect cause an additional voltage-induced power drag on the level of the memory controller. This is important since, as we showed in several earlier articles, the memory controller does eat quite a bit of power and if the same principle is applied to yet another power hog in form of the L3 cache, then one might be able to achieve some significant savings in energy consumption and heat dissipation. Our power measurements show that this is at least worth a discussion. Unfortunately with our methodology it is not possible to dissect core and NB power consumption but the fact that the 9600 at the same core frequency and core supply voltage draws about 25% less power than a 9900 under the same conditions makes a good argument for our claims.

Power and energy efficiency are one side of the coin but what kind of trade-offs have to be expected with respect to performance. We are trying to address this question by repeating the same benchmarks we ran throughout this article on the Phenom 9600 overclocked to 2.6 Ghz. Functionally, the only difference is the NB running at the 1.8GHz frequency as opposed to the 2.0 GHz setting. All benchmarks were run with the memory bus clocked at 1066 MHz and the Turbo button set to red.

The blue bars indicate where the Phenom 9900 performs better than the Phenom 9600, both running at the same clock speed. The red bars indicate where the 9900 performs worse than the 9600. The power numbers show higher (worse) energy consumption by the 9900 compared to the 9600. Across the bench, there is really no difference beyond the margins of systematic measurement errors.

It may be a moot point but you never know until you try. To be on the safe side, we repeated the same benchmarks with the memory clocked at 800 MHz. Again, both CPUs were clocked at identical core frequency, that is 2.6 GHz

In the same fashion as above, the blue bars indicate where the Phenom 9900 performs better than the Phenom 9600, both running at the same clock speed. The red bars indicate where the 9900 performs worse than the 9600. The power numbers show higher (worse) energy consumption by the 9900 compared to the 9600. Again, we fail to see any significant difference in performance.


(AMD Phenom 9600 2.3GHz
(HD9600WCGDBO))

next page: => Green, Yellow, Red... no Rime no Remmon no Cherry =>

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