Urgent.News

What's breaking now, across thousands of outlets.

AI

Nvidia's DLSS 5 Neural Rendering Tech Officially Launches on September 3, Hits Performance by Around 50–60%

Just last week, Nvidia's controversial DLSS 5 technology leaked when a DLL file was found in the NBA 2K27 beta. It shouldn't be too surprising, then, that the technology will be officially launching on September 3 in, well, NBA 2K27.

Nvidia's DLSS 5 Neural Rendering Tech Officially Launches on September 3, Hits Performance by Around 50–60%

Nvidia's controversial DLSS 5 technology officially launches on September 3 in NBA 2K27. The company describes DLSS 5 as a 3D-bound neural rendering technique that uses game geometry and color data to create AI-generated images on top of the rendered frame. This method aims to increase photorealism while avoiding the high computational costs of traditional methods.

Nvidia assures that DLSS 5 won't add extra elements to the scene or misplace objects, as the model is constrained by the original scene's geometry and color information. However, enabling the DLSS 5 neural renderer in NBA 2K27 will likely result in performance drops of around 50-60%.

The company recommends running DLSS 5 in conjunction with other parts of the DLSS software suite, such as super resolution and frame generation, to maintain high frame rates. The technology will only be available on RTX 5000 graphics cards at launch, but Nvidia suggests enabling it only if you have an RTX 5070 or better.

While DLSS 5 has faced controversy, many users believe it improves the visual appearance of games. With its official integration into NBA 2K27, gamers will have a chance to test its performance and visual impact firsthand. The author expresses excitement about seeing how the technology performs and where it may fall short.

Written by urgent.news from IGN's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at ign.com →

More in AI

I raced six models against each other on DigitalOcean Inference. The cheapest one won.

Every time I put a model behind an endpoint I make the same lazy decision. I pick whatever I used last time, or whatever I read about most recently, and I tell myself I'll benchmark it properly later…

  • DigitalOcean's Inference platform benchmarked six different models
  • Integration with DigitalOcean's API was simple and straightforward
  • All six models executed successfully despite inaccurate catalog information

Fambot introduces an ‘AI chief of staff’ for families

Fambot is building an AI “chief of staff” to help families manage the emails, calendars, school updates, sports schedules, and other logistics of raising kids.

  • Fambot introduces AI chief of staff for family management
  • AI agent handles mental load of children's activities
  • Beta testing with 1,000 families shows effectiveness

More from Tuesday 1 September →