Performance
Performance has been measured on our usual DDR4 test rig which is based on Intel Core i7 5820K processor and ASUS Rampage V Extreme motherboard.
I have raised CPU clock to 4.2GHz and cache clock to 3.5GHz so it will be easier to compare memory results. These settings are also easy to achieve on any available Intel X99 motherboard so shouldn’t cause any issues also for our readers.
AIDA64 Memory & Cache benchmark
I’m starting as usual from AIDA64 benchmark which is one of the best tools to compare memory bandwidth and latency.
We can see that with higher memory clock comes higher memory bandwidth but difference between G.Skill DDR4-2800 and Kingston DDR4-3000 is not so big. G.Skill is even faster in memory write bandwidth which usually rely on CPU speed.
I think that numbers in this benchmark are clear enough and don’t need further explanation. I can only add that differences would be lower if we used standard CPU cache clock.
HyperPi 32M
HyperPi 32M is our next benchmark which reacts pretty good on memory performance. In this test we can see that performance of G.Skill DDR4-2666 at CL15 is slightly better than DDR4-2800 CL16. Looks like lower latency helps more than higher memory clock in this test.
Our results are still limited by memory controller’s bandwidth but general performance looks good.
Cinebench R11.5 & R15
In rendering benchmarks like presented Cinebench R11.5 and R15 performance of Ripjaws 4 DDR4-2800 is really good. Difference between memory kits isn’t really big but we clearly see that there are no performance drops and it’s scalling good comparing to other memory kits.
PCMark8
PCMark8 base on popular applications and games. It’s showing us general performance of our computer where we can see how big difference makes memory. It’s hard to say if we notice any special difference between DDR4-2666 or higher memory kits in our comparison but for sure there is some improvement at each higher memory clock.
3DMark Fire Strike Extreme
At the end something which performs 3D tests similar to what we can see in new games so 3DMark. We have used Extreme preset to measure performance of our memory kits as it’s more demanding.
As we see on the above table, memory performance is affecting mainly tests where are available physics calculations. Once again difference between DDR4-2666 or higher memory kits isn’t big and we barely see any difference during test. At the same time we can tell that Ripjaws 4 DDR4-2800 is performing well.
