Nvidia Reflex, here are the first tests

Nvidia Reflex, here are the first tests
When we talk about video games and evaluate their performance we focus mainly on the framerate, probably the most interesting parameter for most gamers. FPS, however, are not the only thing that matters, at least not for professional players: to compete at the highest levels of the eSports landscape it is not only important to play at high FPS, but also to have the lowest possible latency.

To achieve these results, however, pure power is not enough, Nvidia knows it well: for this reason it has worked hard to improve the gaming experience and offer a competitive advantage not only for professionals, but also for ordinary players . An important step in this direction was taken with the announcement of Nvidia Reflex, a technology presented alongside the new RTX 3000 that reduces system latency in compatible games, without affecting performance.

The Reflex SDK allows you to control the times in which the CPU sends the rendering work to the GPU, avoiding creating an unnecessary work queue, in a certain way synchronizing the activities of the CPU and the GPU. The approach is similar to the Ultra Low Latency mode that can be activated by the drivers, but in this case it should be more efficient since the control takes place at the graphics engine level.

In addition, the GPU clock is controlled and increased when necessary to send the frame to the monitor slightly earlier, saving a few milliseconds and thus decreasing latency. Also in this case we are talking about an optimization of the "Prefer Maximum Performance" function that can be activated today by the drivers.

Software support has arrived on compatible games for some time: those of you who play games of Valorant, Call of Duty: Modern Warfare and Warzone, Apex Legends and Fortnite will have already noticed the “Nvidia Reflex ”appeared in the games. However, until now it has not been possible to measure the true impact of this technology, as monitors equipped with Nvidia Reflex Latency Analyzer were not yet available.

Today we finally have not only the software, but also the hardware available. to do this type of test. We therefore decided to carry out a first test of Nvidia Reflex, measuring the latency values ​​with and without the enabled technology. Over the next few days we will carry out further tests, using for example older generation GPUs, lower-end CPUs and so on, in order to understand in which scenarios the technology performs best.

In order to perform the tests c 'specific instrumentation is required: the monitor must be equipped with Reflex Latency Analyzer, also the mouse and the video card must be compatible. As for the monitor, we received an Alienware AW2521H, equipped with a Full HD 360Hz panel, while the mouse is a Logitech G Pro Wireless, on which a custom firmware was installed for the occasion. Below we leave you, as always, the table with the test configuration used.

Intel Core i9-10900K processor (available here) Asus ROG Ryujin 360mm heatsink (available here) RTX 3080 FE / GTX 1660 video card SUPER RAM Corsair Dominator Platinum RGB 16GB DDR4-3600 (available here) Asus ROG Maximus XII Hero Wi-Fi motherboard (available here) Patriot Viper VP4100 SSD (available here) Sharkoon SilentStom 750W power supply (available here) For the tests we will see today we used recent video cards based on Ampere and Turing architecture, but Nvidia Reflex should also be compatible with Pascal (GTX 1000) and Maxwell (GTX 900). We will also test using these cards, mainly to understand how much the new technology developed by Nvidia can improve the gaming experience.

Our tests

In our tests we used three of the compatible games: Valorant, Warzone and Fortnite. We carried out two tests for each title, one in which Reflex is enabled and one in which it is not, and for the moment we have ignored the "on + boost" option, to be used only in cases where the configuration is CPU limited. As previously mentioned, in the next few days we will carry out further tests, one of which will certainly involve this option as well.

We have also noticed changes in latency based on the game area you are in; in our tests we have always chosen the same area both for the test with Reflex active and inactive, in order to have really comparable results.

Test with RTX 3080

Test with RTX 3080 show a reduction in latency of nearly 3ms in Fortnite and around 4ms in CoD: Warzone. We are talking about a percentage difference of 16% and 9% respectively, not exaggerated but still noteworthy values. Different speech for Valorant, where Reflex has significantly less impact and reduces latency by only 0.6ms.

It is important to emphasize that the performance of the video card and the framerate have a direct impact on the effectiveness of Reflex. In Valorant, where the combined i9-10900K and RTX 3080 even exceeds 700 FPS at Full HD resolution, the impact of the technology is minimal; as we will see in the following graph, the situation changes when using less powerful GPUs, such as the GTX 1660 SUPER.

Test with GTX 1660 SUPER

Let's start with Valorant: with the GTX 1660 SUPER the framerate decreases and Reflex manages to guarantee a higher margin for improvement, equal to 3.5ms (18%). The speech made above is also reflected in Fortnite and CoD: Warzone, where Nvidia Reflex reduces system latency by 13ms in the first case (43%) and by 7ms in the second (10%).

Conclusions

From our premilinary tests it seems that Nvidia Reflex is actually able to reduce system latency, with benefits that are mainly noticeable with mid-range video cards. We saw the best result in the test with Fortnite and the GTX 1660 SUPER, where Reflex reduced the latency by 13 milliseconds.

We repeat, the technology is mainly designed for professional players, the only ones able to derive a real benefit from a reduction in latency. In order to make the most of Nvidia Reflex, however, these gamers will have to obtain a compatible mouse and monitor, necessary to be able to measure the aforementioned latency in different scenarios and consequently modify the game / PC settings to minimize it. The choice here falls on a limited number of products: the only monitors currently available equipped with Nvidia Reflex Latency Analyzer are the Dell Alienware AW2521H and the Asus ROG PG259QNR, while for mice it is possible to choose between ROG Chakram Core, Razer DeathAdder V2 Pro, SteelSeries Rival 3 and Logitech G Pro Wireless, on which however a custom firmware must be installed.

At this point we could tell you that to get the most benefit from Reflex you will have to face an out-of-pocket expense of all, considering that we are talking about very expensive monitors. To give you an idea, the ROG model without Reflex Latency Analyzer module has a price of around 800 euros. However, we are faced with a technology designed for professionals, who will probably want (and will be able to) spend large sums for a product that can offer them advantages. Of course, if you are not professional but just passionate, then it is legitimate to ask yourself a few more questions. If you are one of the latter, just think that you can have benefits simply by enabling Reflex in compatible games, but you will never know how much latency will really change by acting on the various parameters. In short, you can trust that the situation is better, but you will never have a certain data.

When asked “Does Nvidia Reflex really work?”, The answer is “yes”, however we must also add a “but”. The system latency actually decreases, but the results are only appreciable with mid-range and (imagine) medium-low range video cards. We hardly imagine a professional playing with a low-end system willing to buy a 360 Hz monitor with integrated Analyzer. In addition to the fact that with a mid-range or low-end configuration it would not even be able to reach a very high frame rate, which everyone is aiming for. And in the case of high-end configurations, the difference Reflex makes is limited. Of course, it is measurable, but are the 3 or 4 ms of change really visible?

There is one point to analyze, which is what we are working on: the influence of Reflex at higher resolutions than Full HD. The only way we have to achieve these results is to use Nvidia's Super Resolution, considering that currently compatible monitors are only Full HD. However we want to make sure of the veracity of the data, which at the moment are fluctuating. However, we believe that this is precisely the key to interpreting this technology, namely the possibility of having latencies similar to those in Full HD, at higher resolutions.

So, in conclusion, the results we have obtained at Full HD resolutions aren't very encouraging, but before we give a final verdict, we'll add higher resolution tests to the rating. In the meantime, have you got an idea of ​​what Nvidia Reflex can offer to the world of gamers? Will it be a key technology for you, or just one more feature that won't make a big difference?





Powered by Blogger.