What is the impact of system memory on gaming performance? It’s a topic we like to come back to periodically, especially when things change. This happened when Intel released the 11th Gen Rocket Lake platform and the idea of ââdecoupling Gear 1 and Gear 2 memories with promises of super high speed RAM.
Compared to the 10th gen cores, the 11th gen platform introduced a fundamental change in the memory subsystem, which decouples the memory controller and the memory frequency.
Some time ago we ran a series of tests with AMD’s Ryzen 5000 series, and came to the conclusion that DDR4-3600, with good timings, offered a good level of performance, and that you shouldn’t. not waste money chasing RAM speed for games. This time around, we’re using Intel’s 11th Gen Core i9 11900K, with its promise of passionate-level memory overclocking performance.
Has that changed, with the potential of 5,000 MHz memory kits, and do we envision a world where high performance memory has a tangible impact on the performance of our gaming machines? Once again, this is a no on our part.
Intel processors traditionally run the memory controller and memory clock at the same speed, or 1: 1 with each other. The 11th generation Rocket Lake processors introduce an option to set the memory controller to half the speed of the memory clock, or 1: 2. Intel calls the 1: 1 mode Gear 1 and the 1: 2 mode Gear 2.
For example, on DDR4-3200 with Gear 1, the memory and memory controller both run at 1600 MHz (remember DDR refers to double data rate). When Gear 2 is selected, the memory speed would be 1600 MHz while the memory controller is reduced to 800 MHz.
RAM: TeamGroup ARGB DDR4-3600 2x 8 GB; G.SKill TGrident Z Royal DDR4-4000 2x 8GB
CPU: Intel Core i9 11900K
Cooler: NZXT X73 360mm all-in-one
Motherboard : Gigabyte Z590 Aorus Tachyon
GPU: MSI RTX 3080 Gaming X Trio
Power supply : Corsair AX1000
Storage: Adata XPG Gammix S70 2TB SSD
Operating system : Windows 10 Professional 64-bit (update 20H2)
Our test system consists of an Intel i9-11900K processor, configured to operate at a fixed frequency of 5.0 GHz. This is to ensure that changes to the memory subsystem are more clearly visible and are not unduly affected by turbo modes popping all over the place.
To date, the Gigabyte card is the only Z590 that allows us to test the particular G.Skill DDR4-4000 Trident X Royal kit that we use at DDR4-4800 with complete stability, although at voltages that we would not recommend. not for a 24/7 user. We also use a set of TeamGroup Xtreem ARGB DDR4-3600.
Both kits use the Samsung B-Die, a favorite with overclockers thanks to their ability to execute tight timings at any given speed. In contrast, most of the brand new DDR4-5000 kits optimized for Rocket Lake use Hynix or Micron ICs. Are they impressive? Yes. Convenient? No. Expensive ? You better believe it.
Coincidentally, the 11th Gen Intel and AMD Ryzen 5000 processors, while architecturally very different, actually behave the same when it comes to memory. Much like Infinity Fabric in 1: 1 mode, Gear 1 works like a charm up to around DDR4-3600, but from there it becomes much more difficult and a switch to Gear 2 is required.
Even with a specially designed OC motherboard, such as our Gigabyte Z590 Aorus Tachyon, Gear 1’s DDR4-4000 did not result in any messages. In our testing, 3,733 MHz was able to start, but it required far too much SA voltage to be worth further testing.
In a perfect world, we would test at many different speeds and timings, and even look at different capabilities. This is just not possible unless a week or 3 without sleep is an option. We started with a DDR4-2400 test. However, this speed is hardly relevant in today’s market, but it is still a common default of the JEDEC system and we have included the results as a benchmark. Running with Gear 2 makes no sense at this speed because you are losing too much performance.
DDR4-3200 to 16-18-18 reflects the kind of kit you will find in many 2021 gaming systems. This speed, along with DDR4-3600, occupies the right price / performance ratio for both Intel and AMD systems.
We tested at 3600 MHz with tight timings so we could see how a high specification 3600 MHz kit at Gear 1 compares to a 4800 MHz kit at Gear 2. We don’t have one of the very latest. 5000 MHz + kits on hand, but our trusty G.Skill 4000 C15 kit is good for 4800 MHz, and it’s still a pretty darn good set of RAM in any language.
We chose to run DDR4-4400 and DDR4-4800 at identical timings only to see the effect of increasing memory controller frequency and clock.
For our analysis, we tested four games. These are the same games we used in our Ryzen memory scan article. These are F1 2020, Horizon Zero Dawn, Metro Exodus and an old favorite, Civilization VI. Choosing a driving game, third person action game, first person shooter, and strategy game means that we have a good mix of game genres and of engines.
We tested at 3 different settings. 1080p at low settings with a GeForce RTX 3080 which puts the performance bottleneck on the CPU and memory subsystems instead of the GPU. This means that changes in speed and synchronization of memory appear much more easily. This is also very relevant information for competitive gamers and users with high refresh rate monitors where FPS maximization is ultimately desirable.
Next is 1080p with maximum graphics settings. 1440p is taking over, but many gamers are using adaptively-synced 1080p displays, and will be for a while. A card like the RTX 3080 can still be bottlenecked at 1080p even with the maximum settings enabled.
4K testing at maximum settings is important to show what effect, if any, Gear 1 and 2 has in limited graphics scenarios. Not everyone runs the latest and greatest GPUs, so this data can be useful for GPU-limited gamers with older or lower performing cards.
The first thing to check are the DDR4-3200 and DDR4-3600 result pairs. This is proof that in games Gear 2 should be totally ignored at speeds where Gear 1 is able to operate.
The only exception is if you’re trying to save a few watts of power, which may be relevant if you’re using a locked-down 65W processor or having throttling issues. Under high graphics load the differences are pretty minimal, but the ability to drop 10% or even 20% of potential performance in the case of F1 2020 means you should use Gear 1 whenever possible.
1080p memory performance
Things get interesting once you get past DDR4-4000 where you have to use Gear 2 to hit those speeds. If you are looking for high FPS then here is a case to be made for very high speed RAM. The question is whether a difference of 5% or so is worth going from a quality 3600 MHz kit, with low syncs at Gear 1, to a very high speed kit fast enough to overcome the trade-off. speed of the memory controller.
The path you choose will depend on the rest of your system; 3200MHz at Gear 1 will suit a Core i5 11400F and RTX 2060 system, but a system with a Core i9 11900K and GeForce RTX 3090 will gain a bit by upgrading to very fast RAM. In fact, calling it “gain” is a bit of an overkill, more about minimizing a potential bottleneck in some high FPS scenarios.
In almost all sane cases, it’s not worth it. But if you’ve got a lambo in the aisle, who cares? Dark.
4K memory performance
On the other end of the spectrum, there’s basically no reason to go for slower memory than DDR4-3000. Reusing a kit from an older system will do the trick with Gear 1, but if you buy a new one then 3,200 MHz should be considered the benchmark, with 3,600 MHz being the sweet spot for performance and price-performance ratio.
A good 3,600 MHz set at Gear 1 offers good low latency and well-balanced performance that will suit everyone from a CS: GO gamer chasing hundreds of FPS, to a multi-screen racing simulator or to a Cyperpunk 2077 player immersed in Night City. Unless you have a cutting edge system to match, or plan to overclock your system until it smokes, there really is very little reason to shell out big bucks for super memory. fast and single-digit performance gains.
The bottom line is that, like a Ryzen-based gaming PC, a Rocket Lake system is best served by a set of DDR4-3600s with decent Gear 1 timings. You need to increase the memory clock by about 1000 MHz to Gear 2 just for the game in performance.
This means that buying memory faster than 3600 MHz, but slower than 4400 MHz, is a waste of money. Switching to the 4600MHz + range is enough to make up for the Gear 2 performance tradeoff, but then you consider paying a lot of money. And at this point, you’re much better off allocating those extra funds to a better processor or GPU.
But then DDR5 is looming, which will change things once again, so we’re going to be doing this all over soon enough …