Breaking Down The Value Chart
Although my critique of Paul's $500 machine and its performance might have seemed a bit harsh, it’s a realistic assessment of a machine on the wrong side of the somewhat bell-shaped curve I like to call performance-value. In PC enthusiast terms, $500 simply isn’t enough money to get a good performance return on your investment.
Here’s how that looks in a chart that compares all of its results to those of the $1000 and $2000 machine, using the $500 PC as a value baseline:
If nothing else, this chart probably makes Paul long for the days when our budget levels were $650, $1300, and $2600.
My $2000 machine sits on the other side of the performance-value curve, costing far more than the point of diminishing returns. Benchmarks that can’t use all six of its cores conspire against it, and the tests that can use its resources reflect that, at two times the price, it only includes 50% more cores than Don's machine. If we halved our CPU budget and spent twice as much on GPU power, we'd start running into another problem: game settings and resolutions unable to take full advantage of the available muscle.
Our entire test suite is designed to reflect the activities of real-world users, and the average performance results show that most of those tasks don't require $2000 machines. As the most practical build in a competition based on practicality, Don’s $1000 machine achieves the best finish in our value chart.
We began this article by wondering whether Don’s machine would set a higher standard at its price point as the zenith of what we were looking for. With only 2% greater value than the $500 machine, it looks like the perfect value point between Paul's machine and mine. Overclocking pushes Don's value lead to 21%, though.
Our ongoing quest for the best performance value is confounded by the fact that the cheapest unlocked Intel processors are currently the Core i5-2500K and Core i5-3570K at $220 and $230, respectively. Until AMD figures out a way to catch up, it looks like we’re forced to face a market where the sweet spot is somewhere around the $1000 Don was originally allotted.
We’ve discussed the constraints of our benchmark suite thoroughly, concluding that any set of tests designed specifically to emphasize the capabilities of a more expensive machine wouldn't represent an adequate number of real-world buyers.
The largest single market we can try to help is the gamers out there looking to play at high resolutions. And so we’ve decided to take another run at the $2000 PC build, shifting focus from CPU to GPU power. In order to overcome the limits of our gaming suite, we're adding the latest high-end 5760x1080 resolution to our tests. You’ll see that alternative, gaming-centric build tomorrow.