Nvidia Readies Deep Learning Anti-Aliasing Debut with The Elder Scrolls Online Update

Marketing image for Nvidia RTX technology.
(Image credit: Nvidia)

Nvidia has been promoting the advantages of its Deep Learning Super Sampling (DLSS) software for graphics improvement since the RTX 20-series launched back in 2018. The GPUs are some of the best graphics cards and the latest Ampere models rank at the top of our GPU benchmarks hierarchy. DLSS uses the embedded Tensor Cores available in Nvidia RTX graphics cards to run artificial intelligence algorithms that upscale and enhance frames from support games, providing better image quality and higher performance. Today, Nvidia announced the next iteration of the technology, which it's calling DLAA (Deep Learning Anti-Aliasing), first supported in The Elder Scrolls Online.

Back in the early days of DLSS 1.0 when the GeForce RTX 20-series GPU made its debut, Nvidia talked about a technology called "DLSS 2x" that was supposed to use the same DLSS algorithm. It would take the same underlying image enhancement technology, but skip the image upscaling. But DLSS 2x never showed up and instead we got DLSS 2.0, which improved the quality and scaling options and boosted performance. Now, three years later, we're finally getting DLSS 2x, built off of DLSS 2.x and with new branding.

Latest Videos FromTom's Hardware
  • RodroX
    I would love to see how this work for games like Euro Truck Simulator 2 and the likes.

    We will see how this work for the ESO.
    Reply
  • setx
    If the game is not GPU-bound then you can run that NN on regular GPU cores without requiring tensor cores and RTX 20/30. But that wouldn't give nV more money be fun, right?
    Reply
  • hotaru.hino
    setx said:
    If the game is not GPU-bound then you can run that NN on regular GPU cores without requiring tensor cores and RTX 20/30. But that wouldn't give nV more money be fun, right?
    Just like you can run RT on shader cores. Clearly spending that tiny bit of die space on hardware we don't actually need is dumb.
    Reply