The presentation yesterday of the monstrous GeForce RTX 3000 with the new Ampere architecture left many noteworthy notes. Not only do we have more powerful NVIDIA graphics than ever at comparatively better prices than their predecessors: remarkable software proposals have also arrived to take advantage of that power in different areas.One of the most striking was NVIDIA Reflex, a set of technologies aimed at fighting against the great enemy of gamers: latency . Many are the elements that influence this latency, and this proposal wants to help detect the culprits of these delays to minimize them significantly.
Victory is measured in milliseconds
When you shoot in games like ‘Fortnite’, the action is not immediately reflected on the screen: the time it takes to see that shot on the monitor depends on several factors and components, each of which introduces a small delay. The sum of all of them is that latency that makes what we do while we play is shown on the monitor with a greater or lesser delay.
That latency can be more or less noticeable by the players, and although it is not so relevant in certain types of games, it is absolutely crucial in competitive games and especially in shooters and FPS like ‘Fortnite’, ‘Valorant’ ‘Apex Legends ‘or’ Call of Duty: Warzone ‘, for example.
As explained by NVIDIA, in games with a high graphic load, what usually happens is that the CPU creates a “rendering queue” in which it orders the tasks that the GPU – which in this way always has work pending – must carry out to display the frames on the monitor.
This allows you to maximize the rate of frames per second that games display, but increases latency because the frames are waiting to be rendered.That’s precisely where NVIDIA Reflex comes in, dynamically adjusting that render queue so that the CPU and GPU are fully in sync . That also helps to relax the workload on the CPU and reduce the latency introduced by other devices (like the mouse) even better.
The NVIDIA Reflex SDK wants to become a tool that allows both developers and users to analyze this latency and then be able to limit it. It is possible to monitor factors such as game or rendering latency in real time, for example, but as seen in the image, there are many elements that affect the final latency of the system . Broadly speaking, the latencies involved are:
- Peripheral latency : it is the time it takes to process the inputs from our keyboard or mouse, for example, and send those events to the PC.
- Game latency : this is the time that the CPU spends processing those inputs and sending a new frame to the GPU to be rendered.
- Render Latency : This is the time from when the frame enters the render queue until the GPU finishes rendering it.
- PC latency : it is the time it takes for a frame to complete that entire journey. Game latency is added to rendering latency.
- Monitor Latency – This is the time it takes for the monitor to display a new image when the GPU finishes rendering that frame.
- System latency : It is the sum of all the previous latencies.
Until now, measuring this latency was only possible with exclusive high-speed cameras, but NVIDIA has integrated this technology as an option in monitors with G-SYNC support graphics and 360 Hz refresh rates. This technology, called Reflex Latency Analyzer.
Several manufacturers such as Acer, ASUS, MSI and Alienware have already prepared monitors with support for this technology . By connecting any mouse to these monitors, we can analyze the latency introduced by both the PC and the monitor, but if we also use a mouse specially designed for this task – here NVIDIA has partnered with manufacturers such as ASUS, Logitech, Razer, and SteelSeries – it will be It is possible to measure the latency of all peripherals separately to obtain an even better analysis of that latency and keys to reduce it to the maximum.
Thanks to the NVIDIA Experience software, we will be able to analyze this latency and also activate the option that allows us to perform a kind of “intelligent overclocking” to reduce this latency: according to NVIDIA, it is possible to reach 50% less latency in games that support this feature. , which of course can pose a sensible competitive advantage in competitive video games graphics problem.
At NVIDIA they actually have an extensive article in which they not only talk about this new technology, but about the different types of latency that exist in competitive online video games. We are not only affected by that latency in our PC / system, but also by others such as network latency (the famous ping).NVIDIA researchers have detected the importance of this latency in competitive gaming, and for example indicate that the difference between a latency of 12 and 20 ms is notable in the performance when it comes to aiming and then shooting in FPS. For NVIDIA ” being good at competitive shooters is much more than having mechanical talents.”
Although latency is not the deciding factor, it certainly helps and for them” higher FPS and lower system latency leads to more frequent target shots.”
With NVIDIA Reflex and its SDK it is possible to apply these improvements to various games, and in its tests with various titles such as ‘Apex Legends’, ‘Destiny 2’, ‘Fortnite’ or ‘Valorant’ the advantage between activating and deactivating it was diverse but notable in many cases for better graphics. The data from NVIDIA is promising and the extensive latency analysis that its engineers have carried out is certainly very interesting, but now it remains to be seen what the real impact of this technology is on gamers. The truth is that the problem exists and that in competitive games, everything adds up Or in this case, subtract.