Today we compare the Ryzen 9 3950X and the Core i9-9900KS in a variety of games with one of G.Skill's highest quality 16 GB memory kits, the Trident Z Neo DDR4-3600 CL14. For those of you who spend $ 600 or $ 750 on a CPU, buying premium DDR4 memory doesn't seem to be a problem, but the main reason we chose this memory was because we gave you a direct one A comparison between AMD and Intel could enable processors with manually tuned timings.
In the early days of AMD's Zen architecture, it was found that memory bandwidth and latency were critical to maximizing the performance of Ryzen processors. By manually setting the secondary and tertiary timing, Ryzen's performance in games can be continuously improved.
In other words, this is a comparison of the top Intel and AMD desktop CPUs that are not part of their HEDT series. Both processors have been optimized for maximum memory performance.
For this comparison between Ryzen 9 3950X and Core i9-9900KS, we tested a total of 18 games with 1080p and a GeForce RTX 2080 Ti. Both CPUs were tested in their standard configuration with XMP for the loaded Trident Z Neo DDR4-3600 CL14 memory. Then we added a second configuration in which the G.Skill save times are set manually. The Core i9 was installed on a Gigabyte Z390 Aorus Ultra, while Ryzen 9 used the Gigabyte X570 Aorus Master.
Before we take a look at the gaming benchmarks, let's take a quick look at the memory bandwidth. As you can see, the timings set only increase the bandwidth by 4% for the 3950X and 3% for the 9900KS.
Memory latency for the 3950X was also reduced by only 5% and 3% for the 9900KS. So based on these numbers, you wouldn't expect a big performance boost for games, but as you will see, this is not the case.
The first game we tested is Hitman 2, where the 9900KS is 21% faster than the 3950X, which is a remarkable performance margin. However, with the optimized memory, the 3950X offers a massive performance increase of 24% from an average of 117 fps to an impressive 145 fps.
While the 9900KS also benefits from optimized storage, not quite as much as Ryzen, we still see a 10% performance boost in this title, which is enough for Intel to keep the lead, albeit a lower margin of 8 %.
Next we have Resident Evil 2 and here the 3950X with the CL14 DDR4-3600 memory does exceptionally well, with the 9900KS being easily displaced both in stock and matched. With this title, the optimized memory has almost no impact on performance. We only see a 3% increase with a 1% low performance.
If we move on to Project Cars 2, we see another slight 3% increase in performance for the 9900KS.
The 3950X achieved a remarkable 11 fps with an 8% increase in frame rate. Not stunning, but it is enough to place AMD ahead of Intel in this title, especially at 1% performance.
Wins in Tom Clancy's Rainbow Six Siege are uneventful for both processors. We see an increase of ~ 3% where Intel offers a 4% increase in performance. However, with a tightly grouped performance of 1%, the experience with both CPUs was identical.
Here are some interesting results in Battlefield V. Optimizing memory with the 9900KS has virtually no performance impact, apart from a 5% improvement in 1% low performance. This is nothing compared to the 21% increase in 1% low power of the 3950X. The average frame rate was also increased by 17%. This was enough to place the 3950X just ahead of the 9900KS, although it had dropped slightly for the 1% low metric. These CPUs have reappeared and enable an identical gaming experience.
When testing with Assassins Creed Odyssey or additional 3 fps, the 9900KS is immediately 5% faster than the 3950X. With the optimized memory, whose margin shrinks to 2.5%, we now see a difference of 1-2 fps, i.e. the same gaming experience with both processors.
If we go ahead we have the results from World of Tanks and here we see an increase of up to 4% over memory optimization, which is pretty pointless for this title. Nor can these numbers be compared to previous World of Tanks results because we use an updated iteration for testing.
We also updated our Metro Exodus test. We are now in a much later section of the game that is more challenging. Interestingly, the Intel CPU, although the 1% low performance is identical between the 3950X and the 9900KS, is significantly faster when comparing the average frame rate, immediately 14% faster and 9% faster if both CPUs use the optimized storage times.
In Fortnite, the Intel Core i9-9900KS shows almost no improvement in the optimized memory. We are talking about only 2-3 fps at well over 100 fps. The 3950X offers a 5 to 10% increase in performance, but even after setting, the 9900KS was still 9% faster compared to 1% low data.
This time, the 9900KS will benefit the most from the optimized memory, since the average frame rate is increased by 5% and the low result by 1% by 14%. Nevertheless, this meant that the 9900KS was only 2.5% faster than the 3950X, the same performance as it was then.
The division 2 performance is somewhat unusual. Both CPUs immediately enabled a 1% low result of 110 fps and still the average frame rate of the 9900KS was still 15% higher. In the past, both CPUs used the optimized timings, although the 3950X achieved a much better result of 1%, and now the 9900KS was only 6% faster compared to the average frame rate.
The optimized DDR4 memory configuration only delivered a 2 to 4% increase in performance for the 9900KS when tested with World War Z. The 3950X was not much better as it only saw a 5-7% increase in performance, although this was enough to reduce the 9900KS performance leading to 4%. Finally, we see again that both CPUs enable frame rates that are so high that the single-digit% margins mean practically nothing.
When testing with F1 2019, we see only slight performance differences in the optimized storage times, although the profit margin was reduced from 6% to 3% for the 9900KS. Another title where the gaming experience is identical to an RTX 2080 Ti even at 1080p.
Ghost Recon Breakpoint does not respond to the optimized memory. Basically, we see no improvement for both CPUs here. This means that the 9900KS was still 6% faster. Pretty terrible optimization in this title, though we have to say less than 100 fps for both CPUs with an RTX 2080 Ti at 1080p.
The same applies to Red Dead Redemption 2 if we have not manually set everything to high. Here we see a performance improvement of up to 5% with the rotated memory, so nothing to talk about. The power range between 9900KS and 3950X is minimal.
After unpacking, the 3950X remains a bit behind in Star Wars Jedi: Fall Order. The 9900KS is 18% faster compared to 1% less power and 12% faster in the average frame rate. The optimized memory doesn't improve the performance of the 9900KS, but does allow an 8 to 14% increase in performance for the 3950X, and now both CPUs offer comparable performance.
In Modern Warfare, there are very small increases in performance if the optimized memory is executed and both CPUs deliver practically the same performance. The 3950X delivered slightly better 1% performance, although margins are so small that you simply won't notice the difference when playing.
Lastly, we have Shadow of the Tomb Raider, and here the 3950X saw a massive 21% increase in 1% low performance with the tuned memory and 17% increase in the average frame rate. In the meantime, Intel saw no improvement in 1% low performance and only an increase in average frame rate of 6%. That said, while the 9900KS was immediately 11% faster, the 3950X with the optimized memory barely managed to take the lead.
The results seemed a little everywhere, sometimes the attuned memory provided strong gains, sometimes these gains were mild, and just as often we saw no gains at all. What seemed clear was the marginal performance difference between the Ryzen 9 3950X and the Core i9-9900KS for gaming. To get a better feel for it, we review the performance of all titles …
After unpacking, the 3950X was 6% slower on average compared to the 9900KS, exactly the same distance between the 3900X and the 9900K when using the slower DDR4-3200 memory (data from a previous review). Then the 3950X with the tuned memory was 4% slower on average, which is practically an identical gaming experience in all modern titles for all practical purposes.
Here's a closer look at the performance difference per game. The 9900KS was 7% faster, which, as we just saw, meant the 3950X was 6% slower. After unpacking, the bigger wins for the 9900KS were seen in Hitman 2, The Division 2, Metro Exodus, Star Wars Jedi Fallen Order, Shadow of the Tomb Raider and Fortnite.
Now that both CPUs are taking advantage of the optimized storage times, the 9900KS was only 4% faster on average, and this time it is World of Tanks, Metro Exodus and Hitman where Intel makes its biggest gains. We see a big change for Shadow of the Tomb Raider. So let's take a closer look at the gains Ryzen has gained from the matched memory.
In the 18 games tested, the 3950X saw a relatively small 6% performance improvement on average compared to optimized DDR4 memory. The only two-digit gains were achieved when benchmarking Hitman 2 and Shadow of the Tomb Raider. In addition, only half of the games tested with the faster memory recorded a performance increase of less than 5%.
If you've ever seen or become part of a discussion about optimized memory and today's mainstream flagship CPUs, the Ryzen 9 3950X and the Core i9-9900KS (or the regular 9900K), these benchmarks compare the performance of stocks and optimized games should give you many arguments.
The stock results were impressive with the Trident Z Neo storage alone. If you used memory with looser timings and more headroom for tuning, at least compared to the out-of-the-box timings, the profits would be more extreme. As mentioned earlier, switching to premium DDR4 memory is probably not a problem if you spend more than $ 500 on your CPU and probably around $ 1,000 on the GPU.
However, if you buy a cheaper CPU like the Ryzen 9 3900X or even one of the Ryzen 7 models, you don't have to invest in premium DDR4 memory. We confirmed this in our Ryzen DDR4 benchmark for 3rd generation memory scaling. Then we used realistic gaming conditions (not 1080p with an RTX 2080 Ti) and the improved memory performance had little or no impact on frame rates.
These tests also confirm that the Core i9-9900K remains a top gaming CPU and is the fastest for the price. However, we wouldn't call it the ultimate solution just because it doesn't consider the performance crown essential.
Compared to the 3900X, which costs roughly the same, it's an average ~ 5% faster for gaming, but the Ryzen 9 has a cooler and is miles faster in heavy core applications, between 20 and 60% faster. We also believe that the extra cores will make the Ryzen more future-proof, but that's probably less of an issue for those who buy now, as you'll likely upgrade in 3 to 4 years.