RTX and DLSS
Alongside the introduction of the new Blackwell-based GPUs, Nvidia had multiple new announcements on the tech side of things with their newest version of DLSS 4 with its addition of Multi Frame Generation. They also have Deep Learning Anti-Aliasing (DLAA) and Transformer-based models to continue to improve the detail and accuracy of the AI-generated frames. So I am excited to see how it all performed. First up though I do have our standard 3Dmark DLSS comparison which I ran on every Nvidia card from the last two generations. For this test, I tested them all at 4k using the performance setting. I ran the same test with DLSS off, and then with DLSS 1, 2, 3, and 4 to compare the performance between them. Adding in the DLSS 4 results from the RTX 5090 and RTX 5080 broke the scaling of our graph and that puts some perspective on the performance difference that you can see when frame generation gets switched up to x4. For the RTX 5080 Founders Edition, it is interesting to see how the base frame rate was 49.11 and with DLSS 1 it jumps up to 88.48 and then up to 121.9 with DLSS 2. DLSS 3 with frame generation was another big jump, up to 169.01 but it is small compared to the 295.06 with DLSS 4. That is a 500% increase. This is a synthetic test that doesn’t change paths at all, so DLSS is going to handle it better than in a standard game but it is a look at the possibilities.
I also wanted to take a look at DLSS 4 performance as well a little more than my initial look at it with 3DMark’s benchmark. For this, I put Nvidias Frameview to the test to run a few benchmarks using the games that currently support DLSS 4. For games with a built-in benchmark, I ran the benchmark but used frameview so we could get the FPS and 1% lows. All of the tests were done at 1080p, 1440p, and 4k with the highest detail settings including the highest RTX settings. A few of these games had to be run on beta copies, namely Cyberpunk 2077, Hogwarts Legacy, Alan Wake 2, and Star Wars Outlaws. Marvel Rivals and Dragon Age both were on the public version but just used NV App’s override.
The first game tested was Dragon Age: The Veilguard and for this one, I did a test with DLSS 3 and 4 as well as with it off on all three resolutions. Like in our 3Dmark tests, this does a great job showing the performance improvements you can see with DLSS 3 or 4. The RTX 5080 FE went from 63 FPS at 4k and up to 145 with DLSS 3, a 130% increase, and then up to 261 FPS with DLSS 4 for a 314% increase. That takes the playable 63 up into nice smooth ultra-high refresh rate smoothness.
For Hogwarts Legacy, I kept things simple with DLSS 4 on or off. At 4k it jumps from 54 FPS up to 286 and even at 1440p it was worth the improvement going from 95 FPS up to 354. The 1% lows aren’t as tight to the average FPS as they were with DLSS off, but they do improve as well.
I tested the RTX 5080 Founders Edition out in Star Wars Outlaws. This was a comparison between the game with DLSS off and with DLSS 4. I saw significant improvements at all three resolutions and with the settings fully cranked (ultra detail, RTX direct lighting on and on ultra, cinematic lens turned on and on ultra) the game struggled at all three resolutions. Even at 1080p, we averaged 55 FPS and at 1440p and 4k it was barely playable. Turning DLSS 4 on however fixed that with 4k going from 18 to 68 FPS (277% increase), 1440p going from 36 to 131(263% increase), and 1080p going from 55 to 200 FPS(263% increase).
Next up was Cyberpunk 2077 and I changed things up here. This game had the option to switch between the Transformer and CNN models so I wanted to see what the performance difference between them was. The new Transformer model makes a significant improvement in the visual quality but as you can see here it does tradeoff performance slightly at all three resolutions. That improvement is worth it even if the tradeoff was a lot bigger than this, but it is important to note. That said, just using DLSS 4 in general turns unplayable performance when running Cyberpunk with every detail and ray tracing setting cranked up into smooth playable performance At 4k you go from 19 FPS up to 149 and that is the worst case situation. At 1440p it goes from 38 to 298 and at 1080p you go from 63 up to 481.
For Marvel Rivals I put the RTX 5080 to the test and compared the results between DLSS off and DLSS 4 on at all three resolutions. Titles like this typically don’t struggle even at the highest resolutions. In this case at 4k its DLSS off result was 79 FPS which is more than playable, but turning DLSS 4 on you see a 267% increase in performance with DLSS 4 going up to 290 FPS. At 1440p you go from 139 to 409 for a 194% increase. But at 1080p the improvement is even better going from 151 up to 545 for a 260% increase. You might not think this is important because you are starting with playable or even smooth performance, but it takes you up into the 290/409/545 range making ultra-high refresh displays possible.
For the last test, I took a look at Alan Wake 2 with DLSS 4 or with frame generation turned off. The game doesn’t give the option to turn off the other DLSS features but frame generation itself offers a huge performance difference and you can see that here. At 4k you go from 28 up to 99 FPS for a 253% increase. At 1440p you go from 58 up to 119 which doubles the performance but we can see we still have the same frame cap issue that is exclusive to the 1440p results, I saw the same with the RTX 5090 testing previously. Then at 1080p, we go from 89 up to 297 a 233 improvement.