We can say that the processor is the second most important hardware after graphics cards in terms of gaming performance. CPU technologies have made significant progress, especially in the last few years, with the heat of competition from Intel and AMD.
From the 6th Gen Core Skylake to the almost 11th Gen Rocket Lake platform, Intel hasn’t made any improvements that would make users happy. With the rapid development of the Zen architecture, AMD has begun to have a greater say in the processor market. This success of the red team was also effective in pushing Intel to the gas.
The number of cores in the latest generation chips has increased significantly. Until a few years ago, we were seeing 4 or maximum 8 core configurations in flagship models. Now let’s come to our main topic, the performance contribution of core counts and cache to games.
Someone on the internet might say that “x amount” of cores is needed to play any game. However, it is not correct to talk about the core numbers directly because it is necessary to look at the overall CPU performance. Also, the design of Intel and AMD processors is very different. Performance scaling may differ based on architecture, processor model, and design.
Intel and AMD
According to testing on Intel-signed processors, performance appears to scale evenly as cores are added, although any game doesn’t require a lot of processing power. On the other hand, AMD’s Zen 3 architecture chips offer closer performances from 6 cores to 16 cores, and few games show big differences.
To cite some tests with the GeForce RTX 3090 at 1080p resolution, we see that the 16-core/32-thread Ryzen 9 5950X is only 5% faster on average than the 6-core, 12-thread 5600X. On the 5800X, which has 8 cores and 16 threads, the difference goes down to ~3%. You might think the resolution is low at this point, but most of the games tested are CPU-intensive ones.
Let’s continue with the blue team. When using 10th Gen Intel Core chips with the same configuration, the result is much different. Carrying 8 cores and 16 threads, the Core i7-10700K is 9% faster than the 6 core and 12 thread Core i5-10600K. Carrying 10 Cores, the 10900K offers 16% better performance compared to the Core i7-10700K, which is a serious number.
To summarize, core numbers are more important on Intel’s side. For example, let’s say you switched from Core i7-8700K to Core i9-10900K. There is a 67% core difference between them. Yes, that’s pretty essential, but there’s more to it than that.
All AMD Zen 3 processors have 32MB of L3 cache per CCD (Core Complex Die). In this case, Ryzen 5 and Ryzen 7 chips carry 32 MB of total cache. In the high-end Ryzen 9 series, on the other hand, 64 MB of cache can be used as the cores are divided into two separate dies. Ryzen chips are made up of core groups called CCDs. AMD usually does the cache configuration on a CCD basis.
Intel CPUs, on the other hand, show a fundamental change in L3 cache capacity depending on the number of cores. The 10th generation 6-core i5 models have 12MB of L3, the 8-core i7s have 16MB and the 10-core i9s have 20MB of cache.
In other words, when we switch from 10600K to 10900K, not only the cores increase by 67%, but the amount of L3 cache also increases at the same rate. For most PC games, processors like the Ryzen 5600X or 10600K are more than enough. So what’s important here? Extra cache or cores?
TechSpot did some testing by disabling the cores on the 10700K and 10900K, locking the operating frequency, ring bus, and memory timings. The tests used the Gigabyte Z590 Aorus Xtreme motherboard, while the ring bus of the three processors was set to x45, that is, they ran at equal speeds at 4.5 GHz. It also used DDR4-3200 CL14 dual channel memory with all primary, secondary and tertiary timings manually configured.
As a matter of fact, only the core and L3 cache capacity difference, which is constant in each model, remained between the processors. Let’s add that the amount of L3 cache does not decrease when cores are disabled. In other words, even with a single core enabled, the 10900K will still run with 20MB of L3 cache.
How Important Are Cores in Gaming Performance?
All tests were performed at 1080p resolution with Radeon RX 6900 XT graphics card. Specifically, the tested game is CPU-intensive.
We’re basically looking at an IPC benchmark, as all chips run at 4.5 GHz. When we look at the results, we see a 15% difference when going from 6 cores to 8 cores. When it is increased from 8 cores to 10, there is a 9% difference. In other words, there is a significant difference of 25% in the transition from 10600K to 10900K.
However, this difference is not only due to the number of cores, but also the amount of cache. Now look at the 8 core data. When you compare the 16MB L3 cache with the 20MB L3 cache, you’ll see that this alone gives a 5% increase in performance, and interestingly, the two extra cores only increase performance by 3%.
The 10700K and 10900K were already close in terms of performance, so what’s the scenario with only 6 cores enabled? This data is very interesting and very similar to what we see in Zen 3 processors. With all 8 cores enabled, the 10700K is 15% faster than the 10600K in this test. The resulting FPS is 429 in one and 494 in the other. As for the important point, 10% of the margin here is due to the higher L3 cache capacity (12MB and 16MB). So the Core i5-10600K would be faster than the 10700K if it had the same 20MB L3 cache as the Core i9-10900K. The Core i9-10900K runs only 6% faster when 6 cores are enabled instead of 10 cores.
To sum up, a 67% increase in cores provides only 6% more performance, while a 67% increase in L3 cache provides 18% more performance. As a matter of fact, we can say that the amount of extra cache is very important, especially in games.
Here, processors with the same architecture have been tested, so the results are reliable. As we saw in Zen 3, if the chips all had the same L3 cache capacity, going beyond 6 cores for games wouldn’t be very useful.
AMD 3D V-Cache Technology
Let’s not go without briefly mentioning the 3D V-Cache technology. Having managed to place an additional SRAM chip vertically on the processor, AMD can thus significantly increase the L3 cache capacity (64 MB), while offering much higher performance values, especially in games. This important technology was first released with the Ryzen 7 5800X3D. Later, Ryzen 7000X3D models went on sale. Intel does not have a comparable technology, so it is best to compare the processors directly with the FPS values it offers.
Conclusion
As a result, we can deduce from this that the amount of cache is very important for games. The scenario will change depending on the game. but in general this is the case. When it comes to Intel processors, more cores means more cache, so we can’t choose which one we prefer more. However, if this were possible, we would have made the choice in favor of the cache.
Regardless of Intel or AMD, the cache often provides high performance gains if the processor is at its limit when playing today’s games. The kernel is less important. For example, you can observe that there are no big differences in games between 6 and 16 cores. Of course, there are games that use cores heavily, especially multiplayer games, but their number is limited.
As an additional note, you cannot determine which is better by comparing the cache capacity of different CPU architectures. In fact, the same goes for kernels. When it comes to CPU performance, there is much more to it than the number of cores; frequency speeds, amount of cache, architectural design, manufacturing technologies and more. That’s why you should evaluate CPUs by their real-world performance, not their paper specs.
