Review - Xabre 600 Pro by PowerColor
Review by James on 04.30.2003
For benchmarking, I used a Chaintek GeForce 4 MX 440 w/ 64 MB RAM as a ‘standard' by which to compare the Xabre 600.
The Test System:
-Athlon 2100XP Palamino at 1733MHz
-512MB Crucial PC2100 RAM
-2 Maxtor 7200RPM 40GB HDDs ATA100 and a Western Digital 7200RPM 100GB ATA100 drive connected via the onboard Promise Controller
-Windows 2000 Pro SP3 Build 2195
-All drivers were updated before testing and used the stable versions when a choice was available except for the graphics card, where I used the performance driver ver. 3.10.58 which is NOT Microsoft WHQL Certified, for anyone who may care.
The engine and memory clocks on the Xabre came set at 300 MHz, but overclocking them was quite simple. The memory was stable clocked up to 325 MHz, which is not a bad overclock for a stock HSF held on with a pull tie. Anything above this immediately produced screen artifacts; thus, cooling may not help improve memory overclocking. The engine was much more durable. It clocked to 350 MHz initially without a problem. However, after about 15 minutes, it started to produce just a blank white screen and then locked. However, at 346 MHz, no problems came up in the entire range of testing. All of the benchmarks were done with the overclocked settings. Going above 350 MHz, text became unreadable and pictures were horizontally distorted. A 15% overclock without having to do anything other than click is not unimpressive. As far as other driver settings are concerned, I went ahead and made ‘normal' changes a gamer would likely make: performance vs quality setting was not readjusted (to maintain a balance) and Vsync was turned off. I enabled the Xmart vision and set the brightness to 2 out of 3. This latter changed made the greatest difference in UT2K3. Whereas without the XmartVision enabled, I could barely see into the dark spaces, the XmartVision made everything visible. Dark spaces simply vanished, and no performance difference was detected.
Before listing a bunch of numbers like they mean everything, I would like to point out an earlier article posted by waxhell a few days ago. The article has its own perspective about the ‘objectivity' which is destroying the reviewers market. Often you see now a bunch of graphs listed without any analysis of the numbers or what the numbers produced because the reviewer is trying to maintain objectivity but in a skewed manner. Now, this isn't necessarily the reviewer's fault, as the people who make benchmarks do not provide adequate statistical tools to properly interpret the data and it SEEMS more objective to do this, but this is not necessarily so. The numbers given to us by the benchmarks are averages, which are greatly affected by extremely small or large numbers in the data that don't conform to the majority of the numbers. Hence, the average says nothing about these outliers and how they might affect our gaming experience, nor does it give an accurate determination of what most of the gaming experience will be like, because it is affected so much by the dropped frames or the easy moments of rendering. Where possible, I have given the lowest frame rates in addition to the average to convey a sense of the above. That aside, here are the numbers:
:: Copyright © 2002-2008 Techware Labs, LLC :: All Rights Reserved