Which $2000 Machine Is Right For You?
The limitations of testing have continuously beset our highest-end builds for the past several System Builder Marathons. Today, we challenged ourselves to uncork the performance of our games using more monitors, higher resolutions, and a graphics subsystem that any enthusiast should envy.
We’re hoping that wins in the game tests are more than adequate to compensate for losses in the threaded benchmarks dominated by the six-core Core i7-3930K processor.
Unfortunately, we fell just short of our goal, though not necessarily because of its premise. Instead, we find the new machine suffering a 3% value loss due to a processor that came up 2.2% short on matching Sandy Bridge-E's overclock. Not impressive, Ivy Bridge. Additionally, we wasted, if you want to call it that, more than $200 on parts to support four-way SLI. Had we dipped in below the budget, today's machine could have picked up an additional 11% on its value score through simple cost savings.
This high-end machine is only slightly limited by its mid-range CPU and cooler. Had we also chosen a mid-range motherboard and power supply, we really would have ended up with a mid-range box with expensive graphics. That whole high-end title has to be worth something at least, right?
Victory is declared when we look exclusively at the purpose for which this system was specified, Surround gaming. If we narrowed down our criteria to only the most aggressive test settings, today's box would have scored an even more decisive finish.
Reality sets in when we remember that our $2000 PCs have to compare against the $500 and $1000 builds at the pathetic resolution of 1280x1024. Core count doesn't matter at that resolution, and even modest graphics cards are taxed much at all. Instead, CPU frequency determines how each machine performs, in much the same way it dictates the outcome of our iTunes and Lame benchmarks. The difference is that owners of all three systems might transcode music, while anyone with a $400 graphics card (or an $800 SLI setup) won't touch 1280x1024, or even 1680x1050 for that matter. Because so much of our benchmark set is targeted towards mainstream components, the competition is rigged to favor middle-priced-and-below complete systems.
We’d like to change that, but doubt that Paul will ever get to pick a parts combo that handles Eyefinity or Surround gracefully. Until that becomes possible, we'll continue reading your feedback for suggestions on ways we can test to better satisfy your curiosity!