Performance
Performance has been tested on the AMD Ryzen platform, including the Ryzen 7 5700G processor, ASUS Strix B550-XE motherboard, ASRock RX6800XT Phantom Gaming graphics card, Silicon Power US70 1TB NVMe SSD, AbkonCore 850W Gold 80+ PSU, and Enermax LIQMAX III 360 CPU cooler.
All results were performed on the FURY Renegade RGB 16GB DDR4-4600 memory kits. Results marked as red include two memory kits, so 32GB. Results marked as blue include a single memory kit, so 16GB. All these results are based on settings that passed longer stability tests and mixed performance tests.
All results except DDR4-4933 run at 1:1 IF and IMC ratio. DDR4-4933 couldn’t pass stability tests as the motherboard is from more standard series. If we used higher ASUS ROG or other top series motherboards, then we could probably reach DDR4-5000. On the other hand, most users won’t aim so high, so our tests should provide results available for most users.
Let’s begin the tests.
Read and write bandwidth in AIDA64 is scaling as expected and reacts well to memory frequency. On the other hand, memory copy suggests the performance in daily work, which also reacts to memory ranks. In some situations, we can expect a similar performance at DDR4-4266 in a dual-rank configuration as DDR4-4733 in a single rank configuration. In other, the higher frequency will be faster regardless of memory ranks.
Both XMP profiles are performing well, but we couldn’t set all four memory modules together at DDR4-4600. Even though Ryzen 5000 APU series has a strong memory controller, then it’s still not enough.
A single memory kit configuration offers lower latency, but all results are pretty good. We have to remember that Ryzen 7 5700G doesn’t offer improved latency as the non-APU series. A much higher synchronous memory controller and infinity fabric ratios give us even more.
PCMark 10 is the first test where we can see a significant difference between two and four memory modules. It affects mainly the Digital Content preset but is clearly visible.
UL benchmarks like 3DMark Time Spy and Fire Strike series don’t react to memory settings much. We can still see a difference, but it wouldn’t be so significant during work or gaming.
In the VRMark, almost only the least demanding Orange Room shows significant changes in performance. We can rarely see over 20FPS difference between various settings.
Final Fantasy XV and Superposition results are surprisingly close at all settings, and slightly better results are on a single memory kit and its higher frequency.
At a lower display resolution, modern games react differently. However, in our tests, the dual-rank configuration is usually faster. This situation can be seen in the Shadow of the Tomb Raider and FarCry 5.
Whether we are using one or two FURY Renegade DDR4-4600 memory kits, the results are high and should satisfy all gamers. Slightly better results are on four memory modules which we can compare to the dual-rank configuration. On the other hand, it’s tough to find dual-rank memory kits running at more than DDR4-4000 and even harder is to set them at higher frequencies. With new processors and stronger memory controllers, it’s possible to improve the performance using faster memory kits.
