EpilogueClever-Power's VGA Booster
& CPU Stabilizer
most definitely aids in smoothing the DC sine wave between your power supply and system, albeit the CPU or graphic cards. To the extent your system can benefit from "line conditioning" will of course depend on the quality of the PSU you own. Neither of these units can compensate if your PSU is under powered, nor can they correct problems which may exist in the circuitry or devices themselves. For example if your system re-boots because your trying to run an H20 cooled, Pentium-D 930 and a pair of X1900XTX in Cross Fire with a 450W PSU, you?re in trouble. Real World Results - Test Methodology
Beyond the oscilloscope bench-tests, I felt it a challenge to exemplify how these units might aid in enhancing system performance, or correcting anomalies. When I originally spoke with Clever-Power representatives I thought the CPU Stabilizer
might aid in reducing random re-boots experienced by Smithfield and Presler Dual Core CPU owners whom had underpowered power supplies. Additionally I hoped the VGA Booster
might aid in re-boots I experienced while attempting to run dual 7950GX2's in QUAD SLI mode, off the XG Duro 900 and the Mushkin 650XP. The tests proved to be more difficult then I had anticipated. Testing
Intel Test System
|CPU ||Pentium-D 930 Retail 3.0GHz Dual Core 2MB x2 L2 (1.285V ~ 1.388Vcore) Socket-775|
|Mainboards ||Asus P5WD2-Premium (BIOS 0606)|
|Memory ||Super Talent PC2-8000 (2x1024MB 4-4-4-12)|
|Graphics ||AOpen Aeolus 7800GTX-DVD256|
|Power Supply || 1.) XG Duro 900|
2.) Mushkin 650XP
|Cooling ||XSPC X20, XSPX Tripple-120 (3x 120mm XINRUILIAN RDH1225S), Alphacool 12V ~ 24V Centrifugal (OASE)|
|Operating System ||Windows XP|
In order to re-create the random re-booting end users were experiencing with Smithfield and Presler CPUs which usually occurred with under-powered PSU's I tested a Pentium D 930 and Asus P5WD2 Premium with several PSU models. Overclocking the Pentium D 930 I chose the Mushkin 650XP. Without raising Vcore the FSB was increased around 270FSB ~ 275FSB at which time re-boots would eventually occur, usually during intensive 3D benchmarks. On a side note I installed the VGA booster at this point which didn't alleviate the problem. The CPU Stabilizer was then placed into the system with slightly better results.
AMD Test System
|CPU ||Althlon64 3800X2 AM2 Retail|
|Mainboard ||Asus M2N32-SLI Dlx|
|Memory ||Crucial Ballistix PC2-8000 (2x1024MB 4-4-4-12)|
|Graphics || Leadtek 7950GX2 (x2)|
|Power Supply ||PCPower&Cooling Turbo Cool 1KW|
|Cooling ||Corsair Nautilus 500 (modified) Viscool V2 CPU waterblock|
|Operating System ||Windows XP|
Next I tested the same system swapping out the Mushkin 650 XP for the XG Duro 900. Surprisingly this PSU was unstable around 265FSB ~ 270FSB and would often "hang" during start-up. Inserting the CPU Stabilizer and/or VGA Booster failed to improve instability.
Testing the VGA Booster involved the AM2 system running the Leadtek 7950GX2 in single and Quad SLI. With the Mushkin 650XP and a single 7950GX2 I was able to run the system at default speeds, inserting either Clever Power device had little influence on performance. Neither the Mushkin 650XP nor XG Duro 900 would run the system with two 7950GX2's. Adding the VGA Booster and/or the CPU Stabilizer had no impact on system re-boots during start-up.
Finally installing PCPower&Cooling Turbo Cool 1KW is where things began to get interesting. Benchmarks runs with the VGA Booster and/or CPU Stabilizer running a single 7950GX2 did show a slight improvement. Due to the fact the nForce drivers I used were beta versions at the time of testing I eschewed Quad SLI setup for a single 7950GX2 and overclocked the AM2 3800X2 from 250FSB and 300FSB, or 2.5GHz and 3.0GHz respectively. In both circumstances Vcore remained at default settings or Auto.
The charts below exemplify my findings as subtle as they we're. I repeated the tests several times to verify performance benefits.
Those familiar with my articles will find I have a propensity to wax philosophical. When you begin to understand the nature of the similarities between the brain and the computer you too may find yourself running Parallel Distributed Processing software instead of Futuremark 3D Benchmarks. As I approached this review I was reminded of difficulties in explaining the processes which occur in proverbial "Black Boxes" albeit from the Computer Science or Neuroscience viewpoint as in Ned Block's essay The Mind as Software of the Brain.
Exploring disciplines such as AI (Artificial Intelligence), Connectionism or Neural Networks involve many of the same investigative techniques. Today we've delved into a literal "Black Box" (or boxes) and while a few of us may have been interested in discovering exactly what's occurring inside the box, most are interested in it's "behavior." That is to say, how the device in question benefits your system and what is the monetary value of that benefit. Perhaps in this review, more then any other I found methodological similarities to Behaviorism and Reductionism.
By now you?re probably wondering what I'm talking about and to be honest I was wondering what exactly Clever Power had sent me. It's not that these products lack merit, they do aid in smoothing the DC sine wave and to a certain extent staring current. The problem is they simply don't excel at either and what they do accomplish would be difficult for the End-user to recognize.
When I was into my Audiophile years I could easily spend between $280 and $620 for a 1-meter pair of RS Audio DUET interconnects (cables which connect the source (CD-transport .etc) to the amp/pre-amp) and if you think that's costly Crystal Clear Audio Cables (CCAC) sells a 1-meter pair of interconnects for $1200 and a 2-meter pair of speaker cable for $1900. Can you hear the difference? Absolutely, if you've invested at least $25,000 into the rest of your system which is relatively budget for an Audiophile system including the source (CD-transport and D/A converter), integrated tube amp, and speakers. Remember it's the cable that carries the signal and you would be surprised how many people might run out and spend $15,000 on a Loth-X JI300 Silabatone integrated tube amplifier with solid silver hand-wound transformers, then use Monster Cables (which are truly garbage in High End Audio) which will deteriorate the sound as if forgot to take your new Magnepan MMG out of they're box.
The same is true in the PC world as your system is only as strong as its weakest link. I left High End Audio when I realized I was no longer listening to music, instead I was listening to the equipment. The question then remains, are you willing to invest in a product whose benefits are obvious on an oscilloscope yet difficult to detect in "real world" applications?
Questions/Comments: forum thread