Sign in with
Sign up | Sign in

Benchmark Results: 3DMark 2001

Tuning Cool'n'Quiet: Maximize Power And Performance, Part 2
By

Pay attention to the red line--it indicates total utilization of all cores. There are only two because we were using an Athlon X2 7750 when we recorded the data for this graph. Why a dual-core processor? There's a very good reason for this. If a benchmark cannot use two cores, it probably won't be able to use three or four.

You'll notice that only one core really gets utilized at any given time throughout the benchmark. The situation will likely stay the same with more than two cores. So, having more cores isn't going to help here. In fact, having too many idle cores is bad. Why? Idle cores still consume some power (an issue Intel aimed to rectify with its Nehalem architecture, capable of almost completely shutting down idle cores). If the application doesn't use the cores, there's really no need for additional compute resources in that title.

How well do the processors we tested fare in this old benchmark? Let's see the results.

Now, let's look at the power consumption during the benchmark.

As we noted earlier, 3DMark 2001 doesn't fully utilize all available cores. It's also more of a graphics benchmark, so the integrated ATI GPU comes into play here. That's the reason we're seeing such a big difference between the Biostar TA790GX 128M (with DDR2-based SidePort memory) and the Gigabyte GA-MA790GP-DS4H and GA-MA790GP-UD4H (both with DDR3 SidePort cache).

There are some interesting points to pull from the results. Power management does result in a performance penalty, though it's hardly noticeable in the real world. Since performance isn't directly comparable here, it's hard to gauge which processor is the most efficient. To make matters worse, the level of performance differs depending on the motherboard used.

React To This Article