Zak Storey takes you behind the scenes of our testing plans for 2020 and beyond

Zak Storey

GOOD-QUALITY TECH JOURNALISM starts with a solid, objective testing platform, there’s no doubt about it. However, in print things are more complicated than in the world of digital. There’s a very fine balancing act that needs to be performed: Resources and budgets are less readily available here, and time and space far more limited. Fixed print deadlines and word counts are incredibly restrictive, and because of that we can’t just throw pages and pages of benchmarks at you, without sacrificing quality content elsewhere; whether that’s the fantastic photography our art division produces, or the intricately woven literary entertainment, knowledge, and opinions of our writers. Finding the perfect balance between it all is key.

Maximum PC’s current set of benchmarks and zero-points have been getting somewhat long in the tooth, and frankly they could really do with a good ol’ refresh. To that end, this time around we’re going to take you on a complete tour of what tests we’re going to be introducing in the coming months, how we’ll be testing our hardware, and what we plan to do with the results once we’re done.

So if you’re curious about where we’re heading, or interested in learning how to benchmark your own machine, then it’s time to turn the page and enter into the wondrous realm of synthetic and real-world tests you need to know about for the next few years.

Proving Processor Performance

AS INTEL CONTINUES to push the limits of 14nm, and Ryzen prefers to do its own thing, we feel that overclocking outside of the world of LN2 is effectively dead for the majority of enthusiasts. For that reason OC benchmarks are out of the window straight away. That’s not to say we won’t try overclocking, rather that it’s something we’ll refer to in the copy. Instead, we will report on what exactly the maximum clock speed achieved is at stock for both single and multi-core performance under load.

For our synthetic tests, Cinebench R20, HWBOT x265 @ 4K, and HEVC’s H.265 Decode @ 1080p are perfect. These three tests deal with rendering, encoding, and decoding. The latter two use different file formats and resolutions to better replicate what most professionals are using in the digital world.

Moving on to our memory tests, and because memory bandwidth is mostly attributed to the number of memory channels supported and memory speed, we’re going to skip that. Instead, what does matter is memory latency, as it gives us an idea of how the core structure is set up. AIDA64 Extreme’s Memory Latency test is perfect for that.