NVIDIA Reflex & Boost vs. AMD Anti-Lag & Radeon Boost under DirectX12 – Image Quality, Performance and System Latency in Practice!

1 - Einführung und Testsystem 2 - Testmethodik und Benchmarkruns 3 - Die Systemlatenz bei niedrigen Einstellungen 4 - Die Systemlatenz bei ultra Einstellungen 5 - Die dynamische Latenzmessung 6 - Zusammenfassung und Fazit

The system latency under dynamic conditions

Let’s go into the final spurt. Persevere is the order of the day! To illustrate the dynamics, here is, how could it be otherwise, a short video. Please note that I recorded the videos for illustrative purposes only. Since I can’t and don’t want to upload infinitely large videos to Igor’s server, these don’t show the respective test in full length and detail!

Mouse movement from left to right etc. Always directed to the container, so you have the wooden box on the left and the corner of the container on the right as a boundary and so you can roughly compare the movement for the test. FYI, I have 1,200 DPI set on the mouse here. So, then 3.2.1 Onward march!

Now what? Again nothing for Radeon Boost! And NVIDIA Reflex loses its usual 4 ms under dynamic conditions. You can check that out on Overwatch, you’ll find the 4ms there too. Since I’ve already tested more than I’ve published, I can tell you, it’s always been like that so far! It’s much more annoying that Radeon Boost doesn’t manage to whip out such a FPS boost like it did in Overwatch, thus reducing latency via lowering frame time. There is the ball on the 5 m line and the goalkeeper is already in the shower. And what does AMD do? Can’t hit the empty net… Well, the English footballers can tell you a thing or two about that too!

NVIDIA LDAT data for the RTX 3060 Ti

Here NVIDIA Reflex also starts to scatter a bit more and the standard deviation goes up. Let’s see what the RX 6700 XT does.

Interesting, because this time the standard deviation is not increased. In a moment, the scattering will probably get a little worse. But let’s take a look at some more details!

While Radeon Boost is able to come out of the woodwork and slightly improve FPS without this time dropping 1% low as extreme as Overwatch, NVIDIA’s FPS drops quite easily. If that doesn’t show up in the Frametimes?

In both cases we see a difference in the frame times. For NVIDIA to the negative and for AMD to the positive. Which can also be measured in latency! But I still don’t quite get it regarding Radeon Boost. For one thing, the kick regarding the FPS is missing. If you take AMD’s Radeon Boost down-scaling with 50%, then the 1440p resolution should dynamically go down to 720p. I have indeed tested this live with AMD. The picture of what Radeon Boost does in terms of dynamic resolution is something that any AMD GPU owner can look at without mouse wiggling. All you have to do is hold down the right shift key and Radeon Boost kicks into full gear. Let me show you. Please have a look at the HUD (overlay) the mini-map, the FPS display etc., they are graphically hit by Radeon Boost. That’s not the case in Overwatch at all!

I once set the same scene in 720p native, with the in-game settings unchanged, and lo and behold the FPS go well over 200. Why Radeon Boost only brings it to about 130 to 136 FPS, AMD couldn’t really answer me yet. This is currently still under investigation! I’ll keep trying. Only statement was, it would be probably everything in the green range, which concerns my measurements and foils. The fact is, and the colleague from AMD was able to see this live, that visually there is no difference between 720p native and Radeon Boost fully enabled. Sorry if I can’t show you comparison slides and videos to 720p native. That would have been just way too much for this article. If you want to see this comparison, then write me in the forum, then I make a separate article about it. Now it’s time for a conclusion….