Motherboard Benchmarks

MOTHERBOARDS ARE WHERE things get surprisingly interesting. We’re seeing manufacturers tweak their boards and push more voltage into certain areas at stock, so it’s vital that we keep a vigilant eye on that. That way you can stay informed and decide whether you want to purchase a motherboard that might give you more stock performance in favor of more heat and a shorter life cycle, or vice versa. To help identify those situations, we’ve included a bevy of benchmarks aimed at every area of system use.

We’re expanding our traditional synthetic rendering performance test, Cinebench R20, to not only include multicore but also single-core performance. While this is running, in a similar manner to our CPU benchmarks, we’re also going to measure max sustained single and multi-core clock speeds. Although we’re not swapping out processors, this will give us an indicator as to the stock autooverclocking performance of the board.

We’ve picked up AIDA64 again for memory latency, and this time bandwidth

too, followed by CrystalDiskMark 7.0 to identify potential problems with PCIe SSDs.

From there we move into some realworld testing, again throwing in a run of our now-staple Adobe Premiere Pro Benchmark @ 4K. We’re not expecting massive differences here with render time, but if the board is auto-overclocking more than it should, and we’re getting higher temperatures and higher synthetic results but not seeing a decrease in render time here, then there’s a problem.

To capitalize on all that we’ve also paired it with an extensive array of power draw and temperature tests. We’ll be using a 10-minute run of Prime95 Blends test to note the max CPU temperature and max VRM temperature, and then using that same run to identify CPU power draw from the wall under load. Finally, we’ll perform one last run on 3DMark: Fire Strike, running a looped “Combined” test for 10 minutes, noting the max wattage drawn from the wall to better simulate intense gaming loads.