Performance
Performance has been tested on the latest AMD Ryzen platform, including the Ryzen 9 5900X processor and ASUS Strix B550I-Gaming motherboard. Some additional tests were performed on MSI MEG X570 Unify, on which photos were taken, but there were problems with overclocking because of early BIOS for the Ryzen 5000 series processors. In this case, to provide results based on the same test rig, we had to use the ASUS motherboard.
Other components are the ASUS TUF RTX3070 OC graphics card, Silicon Power US70 1TB NVMe SSD AbkonCore 850W Gold 80+ PSU, and Enermax Liqfusion 360 CPU cooler. Other components are less important for the final results.
All results were performed on the HyperX Fury 64GB DDR4-3600 memory kit, so this is what you can expect, but of course, overclocking is never guaranteed as it depends on many factors. All these results are based on settings that passed longer stability tests.
Let’s begin the tests.
Memory bandwidth in AIDA64 is scaling well with memory frequency. It doesn’t say which memory setting is the fastest in daily work but shows differences in pure memory bandwidth, which can be later translated into performance gain in some situations. For sure, these results are good and surprisingly higher than those of memory kits at slightly tighter memory timings. It looks like HyperX did a good job with the XMP profiles.
The same if we take a look at memory latency results, results are good. Even at XMP profiles, latency is low.
As usual, rendering benchmarks like the Cinebench series are barely reacting to memory performance. We can still say which memory setting is the fastest, but it won’t really matter if we translate it into daily work.
The most important for most users is a performance in daily usage, which is simulating PCMark 10. This benchmark shows the best results at DDR4-3600 XMP and DDR4-3733 settings.
UL benchmarks like 3DMark Time Spy and Fire Strike series show nearly the same performance at all our settings. The situation is about the same as with Cinebench, where it is hard to tell which setting is really better as all are close to the error margin.
In the VRMark Orange Room, we can see up to 7FPS more, while more demanding tests are not showing any significant differences. The same situation we could see in previous reviews once we moved to a much stronger graphics card.
Final Fantasy XV and Superposition benchmarks are performing well in all our settings. As far as the Superposition shows results close to each other at all settings, then Final Fantasy XV shows a high-performance bump going from DDR4-3000 to DDR4-3600 and a higher memory clock. This is also similar to our previous results where memory settings help gain about 5% higher performance in games.
At a lower display resolution of 1080p, which is still the most popular, new games are already showing up to 19FPS difference. This is only by switching between XMP profiles, while further overclocking is not helping much more. We can say that XMP #1 is optimal for games at 1080p or 1440p.
I feel like I’m repeating myself, but this is one more HyperX review, where XMP settings are offering us high performance, and additional tuning is not really required. As far as XMP #2 is more like a safe option if we had an older or lower series chipset, then XMP #1 gives us high performance in everything we may use. I’m sure that all who decide on this memory kit will be satisfied.
