Last month I received one such e-mail from Asrock that nonetheless caught my attention. It claimed their Z170M OC Formula was the only motherboard to support G.Skill’s Trident Z DDR4-4333 modules. Initially I thought, how useful is that? Are there even any benefits from running DDR4 memory on the LGA1151 platform that high? For the most part we test using DDR4-3000, as it occasionally shows some benefits over the more typical 2400 and 2666 speeds. Going to 4000 MT/s (2000MHz) and beyond is a massive increase in frequency (and cost) and I struggled to imagine where this would be useful, particularly when gaming. Then again, curiosity had gotten the better of me…
So I asked Asrock to kindly send along one of their Z170M OC Formula motherboards. Disappointingly, G.Skill didn’t have any DDR4-4333 memory available and a month later we are yet to see any go on sale, so this news report is appearing more and more like a marketing exercise. However, G.Skill did come back and say they could provide an 8GB kit of their DDR4-4000 memory which is available for purchase. It isn’t the record setting DDR4-4333 memory, but at 4000 MT/s it doesn’t fall far short and will certainly give us a clear indication of whether or not this kind of high frequency memory holds any merit.
Currently there are a few DDR4-4000 memory kits available from the likes of G.Skill, Corsair and GeIL. Of those G.Skill’s TridentZ modules appear to be capable of the best timings at 19-21-21-41 vs. 19-23-23-45 from Corsair, while the GeIL kits are even slacker at 19-25-25-45.
For testing we’ll be using a few select applications and games comparing the Core i7-6700K at various memory speeds ranging from 2133 MT/s up to 4000 MT/s. Helping to maximize gaming performance will be a pair of GeForce GTX 980 Ti graphics cards, if they aren’t able to exploit the potential of DDR4-4000 we fear nothing will be able to. With that said, let’s get down to business.
Test System Specs
Memory Bandwidth Benchmark
Starting at DDR4-2133 we see a throughput of just 20.4GB/s which isn’t bad but less than what we were seeing from the Haswell processors out of the box. Increasing the memory frequency to 2400 MT/s boosted the memory bandwidth by 12% to 22.9GB/s which is typically what we were first seeing from the Haswell processors. Going from 2400 MT/s to 3000 MT/s , the speed which we regularly test at, boosted the memory bandwidth by another 20% to 27.4GB/s. Surprisingly taking the next step to 3600 MT/s boosted performance significantly yet again, this time by another 20% as we hit 33GB/s. Final stop at DDR4-4000 saw the memory bandwidth reach 35.5GB/s making it 8% faster than the 3600 MT/s configuration. While theoretical, the first benchmark shows some promise, shall we go real-world?
