• YEL = baseline system (Dell 745)
• GRE = performance increase from baseline
• RED = performance decrease from baseline

Benchmark Apps
There are a load of comprehensive (and expensive) system testing applications, but I used mostly freeware benchmarks out of cheapness convenience. Using my median system, a Dell Optiplex 745, as an arbitrary baseline, I ran several different applications for the most accurate comparison between systems. Not all of the applications worked where they were supposed to, though. I decided to exclude one of my favorite online hardware comparison tools, Passmark, because it would only run on two of my Windows systems. Also, most of the applications are Windows-only, so the Mac results come from one of the few cross-platform benchmark applications I could find, Geekbench. The XPS 9000′s Geekbench results were measured with its 32-bit app, so the score is probably lower than it should be considering the XPS runs Windows 7 64-bit. Here is a table showing the full breakdown of results.

• YEL = baseline system (Dell 745) / Average
• GRE = performance increase from baseline
• RED = performance decrease from baseline

Findings and Differences
There wasn’t too much here to be surprised about. Even though each benchmark has its own specialized tests with different combinations of hardware, many results fell within a similar range. Outliers seem to result mainly from testing unrecognized or different combinations of hardware. For instance, Brute Benchmark and Novabench didn’t seem to utilize my home server’s dual core CPU when factoring their results. On the other hand, my Dell XPS Gen2 scored an average of 23% below the baseline, but it ranked second (+70%) according to Crystalmark due to a higher emphasis on graphics performance. In fact, looking at Crystalmark’s graphics results alone, my 5-year-old Gen2 performed 180% better than my new XPS 9000.

Disclaimers
With such a wide range of hardware and software, including the countless ways they inter-operate, a true 1:1 comparison gets difficult. Even running the same benchmark software on the same computer can produce inconsistent results depending on the system state. At the same time, it still remains one of the best ways to get a quick overview of system performance along with potential areas ready for upgrades. In the future, I might try to produce even more comparisons by benchmarking individual hardware components. For now, I’m looking into deals on new video cards.