Tech

Call of Duty: Warzone 2.0 CPU and GPU testing

Call of Duty: Warzone 2.0 is a new free-to-play battle royale game, the sequel to 2020’s CoD: Warzone and part of the latest installment of Modern Warfare II. enjoy multiplayer action. Today we’re going to look at the CPU and GPU performance of this game, and as expected, it was a bit of a nightmare to test because it’s exactly what multiplayer games need.

While we usually like to test around 40 GPUs given what’s involved here, it’s just not possible, but we think we have a good sample of current generation hardware. In total, we tested 7 CPUs and 10 GPUs at three resolutions with two quality presets. The results are super interesting which you will see in a second.

But how do you test a multiplayer battle royale like Warzone 2.0? This is never easy, especially if you want to test more than one or two hardware configurations. Collecting accurate data is a real mission that requires a very long time. This is because no two matches play the same way, players take off from different locations, the load on the system can vary, and you also need to land in the same location to complete the path for the reference pass, and ideally you want this do. somewhere where you won’t get killed often, so good luck to you.

We also found that you can actually only complete one pass per game, not only because we got killed at the end of the test almost every time, but also because as the match goes on, the framerate usually increases as there are fewer players. .

For example, when there were only 30 players left, we saw a 25% increase in the 1% lows, although the average frame rate only increased by about 6%, which suggests that the CPU load at the end of the game was much less. We were only able to test this a few times as it required a late game circle.

Speaking of which, we decided to conduct our tests in a central location, as this would allow us to land relatively easily every time, regardless of the flight path and landing zone. A potential downside to benchmarking the quieter part of the map is that performance can be slightly higher than what you typically see in the busier parts of the map, and we looked into this by noticing that performance was often ~10%. better for reference pass.

But there’s not much we can do about it, as you won’t often do even a 20 second pass in a busy area, let alone a 60 second pass. What we did notice is that the scaling was the same for the hardware configurations we tested, so the boundaries between the various CPUs and GPUs we tested should remain the same.

All this means that the hardware configuration testing we did took no less time than all the Modern Warfare 2 testing using the built-in benchmark, but for 44 video cards.

For this testing of Warzone 2.0, a lot of time was lost just waiting for games, eventually loading the game and going to the partition used for testing, and in many cases we were killed during testing, which invalidated the results, forcing us to start over.

For this test, we used 1080p, 1440p, and 4K resolutions using the Ultra and Minimum quality presets. We’ll start with CPU benchmarks that were compiled using the RTX 4090 and we were only able to test the Zen 4 and Raptor Lake CPUs.

AMD Zen 4 processors have been tested with DDR5-6000 memory and Intel Raptor Lake processors with DDR5-6400 memory as these are the optimal kits we have for each platform right now and are similarly priced.

CPU tests

CPU: ultra preset

Starting with “Ultra” preset testing at 1080p, we find the Core i9-13900K to be the fastest processor, overclocking the RTX 4090 to an average of 242fps with 177fps for a 1% low.

It was only 2% faster than the Ryzen 9 7950X when comparing average frame rates, although it was 6% faster when comparing 1% lows.

From 13600K to 13900K, the performance difference is only 7% for the average frame rate and 9% for the 1% minimum, so while the Core i9 is faster, the difference isn’t huge.

Ryzen 7700X, 7900X and 7950X were between 13600K and 13900K. Only the 7600X was slower, trailing the 13600K by 4%, although the 1% lows fell by a marginal margin. All processors allowed the RTX 4090 to display over 200fps on average, which is great to see.

Increasing the resolution to 1440p resulted in a hard GPU bottleneck at around 200fps with the same 1% lows across the board. So even with a GeForce RTX 4090, if you’re playing Warzone 2.0 on ultra settings for some reason, assuming you have a reasonably modern CPU, you’ll almost always be GPU-limited.

Of course this is also true for 4K gameplay as we are running into a serious GPU bottleneck limiting the average framerate with the RTX 4090 to 145fps with a 1% minimum for 77-79fps. In that case, let’s check out the more competitive “Minimum” quality preset.

CPU: minimum preset

For those who want maximum frame rates and a visual competitive edge, the minimum preset is fine, although we’re confident that tweaking some settings will be optimal.

Here again we find that at 1080p the 13900K is the performance leader, delivering an average of 252 fps, making it 4% faster than the Ryzen 7950X. Meanwhile, the Ryzen 9 processors matched 13700K, while the 7700X only matched 13600K. They did it at 227fps, which made both parts 10% slower than the 13900K.

The 7600X is disappointing here at 213fps, making it 6% slower than the 13600K and 15% slower than the 13900K.

Increasing the resolution to 1440p had little to no effect on the performance of the AMD Zen 4 processors, although we do see increases to 13700K and 13600K along with a 1% low at 13900K.

The 13900K remains ahead of the 7950X by a 4% lead, though it’s only a percentage faster when compared to 1% lows. The 7900X and 13700K are now head to head, the 7700X is slightly faster than the 13600K, and the 7600X can match the Core i5 processor.

At 4K resolution, we are primarily limited by the GPU, even at the lowest quality settings. So if you’re playing 4K, Zen 4 and Raptor Lake will provide comparable levels of performance based on what we’re seeing here with the RTX 4090.

GPU benchmarks

GPUs: Ultra Preset

For GPU tests, we have an assortment of GPUs because we weren’t able to test all the graphics cards we wanted to. Using the ultra quality preset at 1080p, it’s no surprise that the RTX 4090 sits at the top of our chart, beating the RTX 4080 by a narrow 12% margin due to mostly CPU-limited testing conditions.

The RTX 4080 was also only 12% faster than the 6950 XT as the Radeon GPU was processing 193 fps, making it 13% faster than the 6800 XT.

As we found in our previous Modern Warfare 2 testing, Radeon GPUs do well in this title, and we see more of that here as the 6800 XT easily beats the RTX 3080 by a 13% margin, while the 6700 XT matches the RTX 3070.

We then see both the Radeon 6600 XT and 5700 XT outperform the RTX 3060 by 20%.

The jump to 1440p gives the GeForce RTX 4090 a bit more breathing room, and the flagship Ada Lovelace is now 24% faster than the 4080 and 48% faster than the 6950 XT.

The Radeon 6800 XT is also now just 7% faster than the RTX 3080, although the 6700 XT was able to outperform the RTX 3070 by a significant margin.

Surprisingly, the 5700 XT delivers similar performance to the RTX 3070, easily outperforming the RTX 3060, this time by a 21% margin.

Then at 4K, the RTX 4090 is 31% faster than the RTX 4080 and nearly 60% faster than the 6950 XT. The RTX 3080 has finally caught up with the 6800 XT and the RTX 3070 has finally caught up with the 6700 XT. The old mid-range Radeon 5700 XT was still about 20% faster than the RTX 3060.

GPUs: Minimum Preset

Now that the minimum quality settings are loaded, please note that the game was completely reset every time you went from ultra to minimum (or vice versa), since a full restart of the game is required to recompile the shaders and load all the settings…

At 1080p, we can see that the Radeon 6950 XT can match the GeForce RTX 4080 and RTX 4090 due to the CPU bottleneck. The 6800 XT also matched the average frame rates of the RTX 40-series GPUs, although the 1% minimum was significantly lower.

This meant that the 6800 XT outperforms the RTX 3080 and the 6700 XT outperforms the RTX 3070. Then we see the 6600 XT is 30% faster than the RTX 3060 and the 5700 XT is 22% faster.

Upping the resolution to 1440p helps reduce CPU bottlenecks, but even so, the RTX 4090 couldn’t beat the 4080, even with the Core i9-13900K on board.

Meanwhile, the Radeon 6000 series outperformed the GeForce 3000 series without issue, with the 6800 XT outperforming the RTX 3080 by 13%.

Finally, we can see that the RTX 4090 is 20% faster than the 4080 at 4K resolution, and the 6950 XT is 12% slower than the 4080. The Radeon 6800 XT is 10% faster than the RTX 3080, the 6700 XT matches the RTX 3070, and in this scenario, the 5700 XT outperformed the RTX 3060 by over 20%.

Wrap

After spending so much time testing Warzone 2.0, we wish we had more data to showcase our efforts, but that’s the way it is with multiplayer games. For GPU testing, the built-in benchmark works well enough that we don’t think the boundaries have changed from our original coverage, even with a few new driver updates.

However, for CPU testing, this really needs to be done in-game, and that’s what we’ll be doing for our big CPU benchmarks in the future. But for Day 1 CPU results, Warzone 2.0 is not viable as any update to the game will most likely break the results, forcing us to refresh all of our data every time. When testing with just a few processors, this is not a problem, but when you are reviewing with more than 20 processor configurations, it is simply not possible.

For the absolute best performance in Warzone 2.0, the recipe is pretty clear right now: Core i9-13900K + GeForce RTX 4090 is the way to go. However, a Core i5-13600K with something like the Radeon RX 6800 XT would be a killer combo considering the 6800 XT can be had for less than $600 today and was generally faster than the RTX 3080.

As we found in our 40+ GPU Modern Warfare 2 benchmark, the old Radeon RX 5700 XT is a weapon in this game, and nothing has changed in Warzone 2. At 1440p at ultra settings, it was only 3% slower than the RTX. 3070 and only 14% slower at minimum quality, averaging over 120fps.

The RTX 3060 is disappointing in these games considering how much it costs, and the Ryzen 5 7600X is disappointing too, though certainly not to the same extent. That said, trailing the 13600K by about 7% was an unfortunate sight, and while we’re sure 7600X owners will be pleased with the performance, the 13600K will outperform here once the GPU restrictions are lifted.

Trade Labels:
  • Nvidia GeForce RTX 4090 on Amazon
  • Nvidia GeForce RTX 4080 on Amazon
  • AMD Radeon RX 6900XT on Amazon
  • AMD Radeon RX 6800XT on Amazon
  • Intel Core i9-13900K on Amazon
  • Intel Core i5-13600K on Amazon
  • AMD Ryzen 9 7900X on Amazon
  • AMD Ryzen 7 7700X on Amazon

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *

Back to top button