Performance
Performance has been tested on the Intel Alder Lake-S platform, including the i9-12900K processor, Gigabyte Aorus Z690 Master motherboard, ASUS Strix RTX3070 LHR 8GB graphics card, Kingston KC3000 2TB NVMe SSD, Sharkoon SilentStorm Cool Zero 850W Gold 80+ PSU, and Noctua NH-U12A chromax.black CPU cooler.
All results were performed on the FURY Beast 32GB DDR5-5200 memory kit. Considering that the motherboard has early BIOS that isn’t tuned well and causes some issues during training, once Gigabyte releases a new BIOS, we can count on better results. At this point, I’m sure that the RAM can run at tighter timings, but not on our motherboard with the current BIOS. The XMP performance is precisely where it should be, and there are no problems with XMP stability.
Let’s begin the tests.
Memory bandwidth in AIDA64 is scaling great with memory frequency. This is what I would expect of the DDR5. I assume we will see 100GB/s+ once higher memory kits appear on the market.
Also, the latency will go down in time, but I guess we can’t complain at about 77ns when DDR5 runs at CL40 or more relaxed timings. We have to remember that the RAM performance is a mix of variables, so we can’t judge the RAM only by main timings or frequency. In this case, memory controller and fast cache also affect overall performance.
PCMark 10 results are already higher than on the last generation of Intel processors. The performance gain is about 10%. This test is essential because it simulates daily work.
There is no significant difference between our memory settings, but the main XMP profile achieves pretty good results.
The same situation is in other UL benchmarks like 3DMark or VRMark.
In the 3DMark tests, it’s hard to tell which setting is the best as some results are close to the error margin. However, in the VRMark, we can see up to 7FPS difference between our settings.
Final Fantasy XV and Superposition results are performing the best on the second XMP profile, which is surprising as this setting is supposed to be the slowest. Depending on the test, one or another setting is the best, so the sweet spot seems to be somewhere in the middle.
I was expecting to see much higher differences in modern games. All three titles are acting about the same, with a 0-2FPS difference. We’ve seen much higher variations on higher and lower DDR4 settings, like in our Fury Renegade review.
Current DDR5 provides excellent performance, and we can only expect it will be better. Right now, it’s not a much better option than the top DDR4 configurations, but it may change once early product issues related to the whole platform are solved, and we see well-tuned BIOS. Of course, DDR5 will be much more desirable once higher frequency memory kits are on the market.
The Fury Beast DDR5-5200 seems like a well-balanced RAM that is fast and offers high stability. It’s hard not to recommend it when we dream about our first PC with DDR5 RAM. It’s also hard to recommend other DDR5 as Kingston is one of not many brands that are actually available in stores.
