Evaluating comparative performance and value
Value summaryWe can normalise performance of the two systems by setting the fastest's result to '100' and then apportioning a figure for the other system(s) based on the relative performance. A table should make this easier to understand
| System name | Wired2Fire Velocity XFire (AMD) | Wired2Fire Hellspawn XFire (Intel) |
|---|---|---|
| PCMark normalised | 83.1 | 100 |
| SYSMark Preview 2007 normalised | 87.6 | 100 |
| 3DMark Vantage normalised | 98 | 100 |
| Average normalisation | 89.57 |
100 |
| Price | £1,109 |
£1,229 |
| Value for money rating (marks per £1,000) | 80.76 |
81.36 |
For example, the Hellspawn's 6916 score in PCMark is the leading result. It's given a metric of 100. The Velocity scores 5478, or 83.1 per cent of the Hellspawn. Similarly, the Velocity manages 87.6 per cent and 98 per cent of the Hellspawn's performance in SYSMark 2007 Preview and 3DMark Vantage, respectively, as shown in the table.
We then derive an averaged normalised figure based on the three results, with the leading system set to 100 at all times. In effect, we're looking at the percentage performance of any system when the fastest system is locked to 100. According to the benchmarks, the AMD-powered Velocity is benchmarks at around 90 per cent of the Intel Hellspawn, based on the results from the three benchmarks. 3DMark Vantage, being a GPU-bound test, brings up the AMD score.
Of course, one could argue that the Velocity could have a £100 more-expensive graphics card, to compete on an equal price footing, and the normalised results would be different. Thus we've added the asking price into account and then derived a value metric. Yes, we could refine the metric and take different results, apportion different weights to the benchmarks - indeed, use different applications - but it provides a back-of-the-envelope method of absolute performance and relative value-for-money.
What we see is that the Wired2Fire Hellspawn XFire is the faster system, no doubt, winning all the benchmarks, but take value into account and the AMD system's not a bad deal as it should provide near-identical gaming speed once the settings - 1,920x1,200, ideally - make it GPU-bound.