After a long stagnation in the processor market, AMD was able to return to the top segment. Recently we reviewed its new Ryzen 7 1800X processor , which in a number of applications managed to make a serious competition for older Intel Core i7. But a simple home user is also interested in gaming performance. This article is devoted to this article.
It’s no secret that the previous generation of AMD FX processors in gaming tasks looked rather weak on the background of Intel solutions with fewer cores. Now the situation can change, which is connected both with better support for multithreading in new games, and with the new architecture of Zen. For this gaming test, one of the older processors for the LGA1151 platform is chosen as a competitor from Intel. It’s a Core i7-7700K with four physical cores, which processes eight threads. Considered AMD Ryzen 7 1800X has eight modules for 16 streams, although it works at lower frequencies. Just a couple of years ago, such a skewing in the number of cores / threads would not embarrass the players who prefer Intel solutions. How will the confrontation of such different processors develop now? The answer to this will give our testing!
Processors AMD Ryzen 7 1800X and Intel Core i7-7700K will be tested in nominal and in overclocking. At the same time, we will try to evaluate the potential of AMD’s younger solutions, created artificially in our conditions. The sophisticated boost-mode algorithm, which in each separate application will differently adjust the core frequencies depending on the load, does not allow to obtain a complete analog of the low-end CPUs by reducing the multipliers on the AMD Ryzen 7 1800X. But after raising the frequencies and at a fixed value, we can estimate the potential of other AMD models in the overclocking. At the same time we will study the influence of the number of cores on the overall productivity.
For the Ryzen 7 1800X, the closest neighbor with a smaller number of cores is the Ryzen 5 1600X, which operates at a similar base frequency of 3.6 GHz with a maximum boost value of 4 GHz. Replace the overclocked Ryzen 5 1600X at 4 GHz with our AMD Ryzen 7 1800X with a 3 + 3 core configuration. Also interesting are the quad-core versions of Ryzen. We get an analog of the younger AMD Ryzen 5 1400. It has 2 + 2 cores with a L3 cache of up to 8 MB. Of Ryzen 7 1800X, this combination can not be directly obtained, but it is possible to include a 4 + 0 configuration, which will provide the same amount of L3 cache. And at a frequency of 4 GHz it will be possible to estimate the potential of such a processor during overclocking. Therefore, we assume that the 4 + 0 configuration will replace the accelerated Ryzen 5 1400. For a better understanding of the situation, we give a table with the standard characteristics of the Ryzen processors.
The configuration of the test system for Ryzen is as follows:
Based on previous tests, Ryzen selected the optimal mode for performance with the kernel turned off. In the overclocking, the processor frequency was fixed at 4 GHz with the DDR4-3200 memory mode (16-17-17-39). With the same parameters, overclocking to 4 GHz involves two more configurations with different number of cores:
The configuration of the test stand is as follows:
In the overclocking, the Intel processor worked at a frequency of 4.6 GHz in combination with DDR4-3200 memory (16-17-17-39).
On all test configurations, the reference GeForce GTX 1080 graphics card was used. Its frequencies are at the standard level, but the fans speed is increased in order to ensure a low temperature regime and to avoid boost-frequency subsidence due to exceeding the temperature limit.
The test modes were chosen to estimate the practical difference in processor performance, taking into account that the potential user will use high quality graphics. Each game is tested in several modes, resolutions 1920×1080 and 2560×1440 are involved. In the most demanding applications only Full HD resolution is involved, but with different graphics settings. All the nuances are described in more detail below. To measure fps in manual mode, we used the Fraps and Mirillis Action!
Test Fire Strike from the test package 3DMark was run at a resolution of 1920×1080. The summary diagrams show data on the total score on different systems and the data of the physics test.
The new benchmark under DirectX 12 was run with the default settings. The total score and CPU Score are shown.
The testing was carried out in two resolutions: 1920×1080 and 2560×1080 with Ultra-quality (including MSAA 4x).
Based on the test sequence of actions, which we use when comparing video cards. But the test is shortened, at the same time, shooting and destruction with the help of grenades of concrete elements were added to it to introduce the element of calculation of complex physics into the test sequence.
For each mode, two applications run for seven repetitions.
Testing was conducted by replaying a small game episode at the beginning of the “Cape Helles” quest, where Gallipoli’s assault is played out. Fighters landed on the shore under heavy artillery fire, and with abundant explosions, performance sags more than most other game moments. There were two test runs per mode for seven repetitions.
Video memory limitation is disabled. Testing is carried out in DirectX 11 using Fraps.
The test modes are three:
In addition, testing in network mode with Ultra-graphics in 1920×1080 was carried out. The frame rate was measured during the introductory scene to the Monte Grappa map of Operation Iron Walls. During this entry, the camera flies over the location. Since the network game depends on a variety of conditions, the spread of the results is very high. Therefore, at least 20 replicas were performed, and the processors were tested only in nominal value.
The testing was carried out with very high and maximum quality graphics (Very High and Ultra) in 1920×1080. Both modes when rendering in DirectX 12, which provided some advantage over DirectX 11.
The standard gaming benchmark was run at least seven times for two test modes.
Performance was measured using Fraps while replaying a small sequence of actions in the port of Carnaca.
The testing was conducted in three modes:
In all cases, active shading of HBAO + is turned off and the adaptive resolution is disabled. For each mode, two game runs were performed for seven test repetitions.
Testing was conducted with the help of Fraps. The basis is taken of our test sequence for comparing video cards, but the duration of the walk is slightly increased and shooting at the end is added. At least seven repetitions and additional starts in case of a wide spread of results.
The profile of the maximum quality of the graphics is selected, the shading of HBAO + is additionally involved. Testing is carried out in two modes: 1920×1080 and 2560×1440.
The testing was carried out with the help of the built-in checkmark with the Very High quality profile. At least seven repetitions.
Two test modes: in the resolution of 1920×1080 and in 2560×1440.
We used the built-in gaming bacchnmark, which was run at least seven times.
The maximum quality of graphics (Ultra profile) is selected, all additional parameters under DirectX 12 are active. Two test modes in different resolutions: 1920×1080 and 2560×1440.
To test used built-in benchmark. Five repetitions. For a comprehensive assessment, the average fps was calculated based on the results of all the test scenes.
All the main graphics settings are at the maximum, additional parameters are active – Extended Distance Scaling and “Extended Shadows Distance” + 100% to the base level. Three test modes with different anti-aliasing settings and in different resolutions:
Instead of the standard beechmark in this game, the performance was measured during the introductory scene to the initial mission. This is one of the moments with a lot of NPCs, which are the most critical for performance in the game. There were no less than seven repetitions of the test scene below for each mode.
Testing was carried out with the maximum graphics settings with anti-aliasing SMAA. At the moment the game began to work faster in DirectX 12, so we decided to compare the processors with different DirectX. There are only three test modes:
To measure fps, we used Fraps and Mirillis Action!
The built-in benchmark was used to measure the average time of artificial intelligence (AI benchmark). The only test where a lower value is the best result.
The graphics settings for the benchmark AI are not critical, but it was decided to take measurements at 1920×1080 and 2560×1440. Both modes set the maximum quality of graphics when smoothing MSAA 4x in DirectX 12.
The testing was conducted with the help of Fraps at the beginning of the first mission after the prologue. A certain sequence of actions was performed on a large location with a complex terrain. At least seven repetitions for each mode.
The maximum quality of all graphic parameters is selected, including “insane” texture quality. Additionally, Adaptive Resolution is disabled. Testing is carried out in two modes at different resolutions: 1920×1080 and 2560×1440.
Testing was conducted with the help of Fraps. Measured fps during a trip on the road to the village of White Garden. At least six repetitions and an additional test session for the simplest graphics mode.
Maximum graphics settings are used. In the case of the simplest mode, “beyond the limit” quality is combined with the disabling of the HairWorks, and post-effects are active only by smoothing, glowing and light columns (beams). There are three test modes:
For testing, a built-in gaming benchmark is used, which was run 7 times.
The game is extremely demanding on the power of the graphics accelerator. Therefore, two modes are selected at the same resolution:
The built-in performance test was run at least seven times for each mode. Testing was carried out only in DirectX 11.
Three test modes:
Ultra + means an additional increase in all parameters to the limit (reflection, detailing), plus a better shading of HBAO +.
The built-in performance test was run at least six times for each mode. Testing was carried out only in DirectX 11, because here DirectX 12 gave only the fall of fps.
Three test modes:
Under Ultra is meant a standard quality profile with the same name. In Ultra + we further increased to the limit all available parameters and activated the heavy anti-aliasing of MSAA.
Running through the streets of the Palo Alto area and the nearby grove. There were two runs of the game for five or six repetitions of the test walk.
Three test modes:
Making measurements of the total energy consumption of the system in automated tests and beechmarks at a resolution of 1920×1080 with maximum graphics quality settings. List of test applications:
Peak values were taken into account during each run, on the basis of which the average peak for the individual tests was calculated, and then the final mean value was calculated. The data was captured using the Cost Control 3000 instrument.
Let’s begin the study of the results with the overall score in Fire Strike.
The final result for the Ryzen 7 1800X is about 4% lower compared to the Core i7-7700K. Overclocking to stable 4 GHz just allows you to compensate for this difference. It is noteworthy that in a configuration with 4 cores (8 threads), Ryzen is weaker than the full version at a lower frequency.
More clear results on pure CPU performance will give a comparison of CPU Score.
And here everything is different. Although in the overall standings, the AMD processor also lost, but in the processor test is far ahead with an impressive margin relative to the Core i7-7700K. Overclocking gives the Ryzen 7 1800X an acceleration of more than 8%. At one frequency of 4 GHz, the hypothetical Ryzen 5 1600X is weaker than the older version by 13.3%, and the theoretical Ryzen 5 1400 is weaker by 55%.
In the newer 3DMark Time Spy, the AMD processor is initially the leader. With him, the total score is 5% higher than the competitor’s. Acceleration gives an additional slight acceleration. The configuration with 6 cores at 4 GHz is at the level of the forced Core i7-7700K to 4.6 GHz. Theoretical Ryzen 5 1400 is weaker than all, but the lag behind Intel’s representative is small.
A pure physics test is demonstrated by a physics test.
Here AMD Ryzen 7 1800X does not leave a single chance of Intel Core i7-7700K. The gap in the physical test at nominal frequencies is more than 45%. The theoretical Ryzen 5 1600X in overclocking to 4 GHz is faster than the forced Core i7-7700K.
Let’s move on to real gaming applications. Let’s start with the old Battlefield 4 .
The frame rate is 150 fps, and the actual difference in results is scanty. A notable advantage of the Intel system, but it is insignificant. It is also worth noting that overclocking does not give any noticeable acceleration. Initially, both processors provide extremely high performance, which rests only on the potential of the video card.
The transition to a higher resolution blurs the differences between competitors. There is a slight advantage of AMD in one frame at the minimum fps, but it’s all on an insignificant level.
In Battlefield 1 , the situation is slightly different. The overall performance even with high quality graphics is lower, the advantage of the Core i7-7700K over the Ryzen 7 1800X is about 4%. Overclocking gives a minimal increase in performance. Ryzen with 4 cores at a frequency of 4 GHz is slightly weaker than comrades, but this is also hardly noticeable.
The transition to maximum quality and higher resolution does not erase the differences between competitors. Intel retains the overall advantage of up to 4%.
As a supplement, consider the results of testing in a network game. Measuring performance in this mode is extremely difficult, because you can not simulate the same events. Each battle develops in different ways, different players’ activity, different weather conditions on the map. Compensate for all this is possible only with a large number of repetitions. This was done by repeating the opening time of the operation on the map of the Monte Grappa map, when the camera flies over the game location. Due to the large time costs for such testing, we confine ourselves to a comparison in nominal modes.
In multiplayer, the difference is more than 8% in favor of Intel Core i7-7700K. But still it is necessary to take into account the factor of possible error. The main thing is that the overall relationship between the competitors are still small.
Game Mankind Divided known stringent requirements for a graphics accelerator. But also a powerful processor provides clear advantages. At very high settings is very good advantage for the Intel Core i7-7700K.
With an increase in the graphics settings to limit processor Ryzen 7 1800X is no longer the weak link, although the lag in a couple of percent retained. A slight theoretical lag processor Ryzen May 1400 can be associated with half the volume of L3 cache.
Before considering the testing results Dishonored 2 , we note that the minimum fps here can significantly ride. So when evaluating this option, it is necessary to take into account the possible error.
Traditionally Ryzen 7 1800X inferior competitor, but when accelerating to 4 GHz Core i7-7700K catches in the nominal. When a single frequency, we see quite a clear variation in the results from AMD. It is unlikely that there is a direct correlation of the number of cores, since the improvement of the quality difference between the 6 and 8 cores scanty. But in the minimum configuration of the nuclei and a marked drop in serious, it manifests itself as a more volatile performance and reducing the overall totals. It may be associated with fewer cache in processor theoretical Ryzen May 1400.
Increasing the graphics settings reduces the gap between Ryzen 7 1800X and Core i7-7700K, but the second retains the advantage. The lowest results in analog Ryzen disbanded in May 1400.
7 1800X Ryzen losing competitor in Fallout 4 . And even after the AMD overclocking weaker competitor in the denomination. However, with the original restriction of 60 frames you the difference in the game does not feel. Cache and the kernel does not affect the results Ryzen, only frequency is important.
Transition to 2560×1440 eliminates all differences. In fact, we have the same results for all processors.
The game For Honor we see the same results. Even in the mode of Full HD frame rate at the level of 135-137 fps difference between AMD and Intel is negligible. So here any processor will be enough to head.
The results in the benchmark Gears of War 4 looks very sad for AMD Ryzen 7 1800X – a huge gap from the Intel Core i7-7700K. The advantage is that with AMD you still get above 60 fps.
Built-in test provides extensive performance data, allowing you to directly compare the results of processors.
And according to these data the difference is smaller but still serious. Also visible to a certain scalability in accordance with the number of nuclei. Theoretical Ryzen May 1400 in the dispersal of less Ryzen 7 1800X in the denomination.
The resolution increases, the difference between processors is not so enormous, but still high. It is noteworthy that the results Ryzen almost no different in both resolutions, AMD processor then acts as a limiter performance.
Another game is a processor and a noticeable lag AMD Ryzen 7 1800X. Acceleration provides a performance boost of about 10%, but against an opponent it does not matter. There is a tendency to lower results in reducing the number of active threads from 16 to 12. The lowest rates in the theoretical CPU Ryzen May 1400, which has the smallest cache among AMD solutions.
The transition to a better smoothing of the MSAA reduces the difference between AMD and Intel processors, but the position of Core i7 is still undeniable.
Influence of CPU on the overall performance can be seen even in high resolution 2560×1440. Then the processor has a greater effect on the average result, the minimum fps difference is very small.
The first results show a serious gap Hitman Intel processor, AMD newbie lag at 30%. However, activation of DirectX 12 somewhat changes the situation.
There is a general increase in the results, and it is higher than AMD. Core i7-7700K still retains the lead, but Ryzen 7 1800X is now inferior to 16-19%. In DirectX 12 processor resources are used efficiently. Comparison of different variants Ryzen at 4 GHz shows a small loss in going from 8 to cores 6 and a perceived decline in performance with embodiment 4 nuclei and trimmed cache (analog Ryzen May 1400).
It is worth mentioning that in the initial stages of DirectX 12 only provided a serious decline in performance in the game. But for the year developers Hitman and the new NVIDIA video drivers were able to improve the situation.
The high-resolution when rendering under DirectX 12, the difference between the processors at the level of a few percent. Gain frequencies in this mode does not provide any appreciable acceleration for Ryzen, but cuts cache Ryzen May 1400 analogue affects markedly.
Benchmark AI is independent of graphics options, and an opponent’s turn in Civilization equally in both resolutions. Intel Core i7-7700K processor copes with this task faster, which will provide a smaller waiting period during the course of the computer opponent on the map. No difference between Ryzen 6 and 8 cores. And here is analogous to an accelerated Ryzen May 1400 is not very different from Ryzen 7 1800X in the denomination.
In Titanfall 2 possible to ascertain the approximate equality of competition. Of course, the overall trend is noticeable with the advantage of Core i7-7700K, but even with an average frame rate of 120 fps for this difference is so small, that does not matter.
The difference between AMD Ryzen 7 1800X and Intel Core i7-7700K at some 2-3%. Such a lag AMD processor easily offset acceleration.
Increasing the settings to the limit does not change the overall balance. Perhaps in some game situations CPU impact is more pronounced, the senior Ryzen definitely will play in the Witcher 3 with full comfort at maximum performance, limited only by your video card capability.
At Very High settings noticeable advantage in a few percent of the processor Core i7-7700K. It is noteworthy that the acceleration does not affect the results of Intel and AMD, as without changing the overall balance.
Increasing the graphics settings to maximum slightly reduces the difference in performance, but Intel representative still shows a definite advantage.
Built-in benchmark The Division considers only average performance. We further present data on the minimum fps, but with the caveat about the large variation of this parameter and the possible errors. So focus on the average fps, as a more accurate criterion for evaluation.
The first results are not at the highest quality graphics show the advantage of living Core i7-7700K of a few percent. Is not critical, but AMD Ryzen 7 1800X even after the dissolution of weaker competitors.
Let us now compare the processors with the highest quality graphics in DirectX 11 and DirectX 12.
Maintain stability advantage of Intel, but it is very modest. We should also note a slight increase in productivity during the transition to the new API, although very little.
The maximum resolution of tiny difference between the processors.
With the standard Ultra-profile settings, the difference between the processors can be clearly seen. AMD Ryzen 7 1800X again inferior opponent, although the overall frame rate far beyond the 120 fps.
Increasing the graphics settings eliminates differences and equalizes competitors. However, this concerns a game mode of real time. miscalculation of the running time on the global map of the processor depends stronger, but no benchmark for measuring this parameter in the game.
According to the monitoring means Watch Dogs 2 excellent zapruzhaet all 16 flows over 70%. But Ryzen 7 1800X still loses to a competitor. Core i7-7700K with its 8 flows at a high frequency is more efficient, providing gain 14-19% in the test stage.
Increasing the graphics settings to Ultra-expected level reduces the gap, but only up to 8%.
The high-resolution (temporal filtering is enabled), the difference is even smaller, but still can see a clear advantage of Intel Core i7-7700K.
The following data allow us to estimate the average power consumption of the entire system at the gaming load.
Almost identical performance from processors in the denomination, and a serious difference in the dispersal. AMD has power consumption with an increase in the frequency and the voltage increases more significantly. However, Intel itself crackdown on the initial frequency miserable.
According to the comparison results in most games we can see a slight difference in the results between the processors AMD Ryzen 7 1800X and Intel Core i7-7700K. No surprises in the form of AMD is no advantage, except for CPU tests 3DMark. In real Core i7-7700K games it looks interesting, but it is worth recalling that this is the record for those frequencies among serial Intel models, and frequency advantage of still playing an important role. But even against such a rival processor AMD Ryzen 7 1800X looks very good. Note the minimum difference in the results in 1 Battlefield or The Witcher 3. Yes, and in many other games when the Ultra-quality difference between AMD and Intel systems at the level of two to four percent, which is not critical. Against the background of the previous generation of AMD processors is a major achievement. We have a processor that is not much worse than the current Core i7 gaming and demonstrates an advantage in the application of multi-threaded tasks.
It is worth noting an interesting trend with good growth performance during the transition from DirectX 11 to DirectX 12 in Hitman. Perhaps the best new API unlocks the power of AMD Ryzen. On the other hand, there are very sad for AMD results in Gears of War 4, where a serious gap Core i7-7700K is observed in the same DirectX 12. Among the poor results may be noted and the lag in GTA 5. In principle, AMD Ryzen 7 allows quite comfortable play the same Grand Theft Auto, but Core i7-7700K frame rate is clearly better. The important role played by optimization. AMD Ryzen – it is a fresh product on a new architecture, and at the expense of program optimization of its results can be a little grow. As an example it is worth mentioning the strategy of Total War: Warhammer – a few days ago it was announced that a special update improves performance on AMD Ryzen up to 10%. Do not expect that all game developers will rush to optimize the game for the specifics AMD Ryzen, but some updates for major projects can go. And in the future, AMD Ryzen support will improve.
It is possible to draw general conclusions about AMD Ryzen 5 1600X potential. Disabling 2 cores (4 threads) are not important for the gaming performance. Therefore, when the same frequencies AMD Ryzen 5 1600X and 1800X AMD Ryzen 7 are as close as possible. And it makes AMD Ryzen 7 1600X better buy for those users whose tasks are limited to the gaming area. But our analogue Ryzen May 1400 was not just a weaker older versions at the same frequency, but also weaker AMD Ryzen 7 1800X with lower nominal frequency. It affects a smaller amount of L3 cache. It should also be noted that the real processors Ryzen May 1400 a series of work in 2 + 2 configuration and the initial frequency of their very low. So without overclocking performance of such solutions can be markedly weaker. But if you can squeeze 4 GHz, the final performance will be close to Ryzen 7 1800X in the denomination.