What are the Nanite console-specific gains we could expect to see from 5.7 and in the future?

Development Rendering
, ,
1

What are the Nanite console-specific gains we could expect to see from 5.7 and in the future?

5.7 Public Roadmap lists a few new Nanite features, but are there any new optimizations made or being actively worked on with an estimated release window?

2

Steps to Reproduce

3

Hi Michael, many of the console-specific performance optimizations for Nanite in 5.7 have been derived from our efforts on the Witcher 4 demo. This means that you likely see the most performance gains if you are going to experiment with our new foliage tech. For Nanite, we typically take a console-first approach when building new features or making optimizations, so you should naturally experience improved performance from release to release. Are there any specific details you are looking for? Thatt would help in focusing the conversation.

4

Hello,

I’m Stephen Johnson, lead technical artist. I’ll try and pitch in with some info. We’re near the start of production and just trying to get our budgets and expectations correct. We will be updating to 5.7, so any improvements there will be gladly received! The question would be more around how complete would you say Nanite is currently? Is there more perf to be had, particularly on low end machines, or is the feature complete now and any gains from now on will be minor? Foliage won’t be a big deal for us, it’ll be more about scene density, more than anything.

thanks

5

Hi Stephen,

Sorry, I was out of town on vacation last week, so I’m just getting back to this now. If I understand your question correctly, Nanite is robust enough at this point to be used on gen-9 consoles and above. Targeting those platforms has always been part of the feature’s development, and we continually make improvements to the system over time. For the 5.7 release, we are adding a few performance and stability improvements, some of which I listed below:

  • XSX-only build option to include Nanite fallback meshes in cooked builds (with per‑asset override in Static Mesh Editor)
  • The work graph computes path safety fixes to prevent GPU hangs
  • Platform-aware base pass shading uses Work Graph shaders when available (common on gen-9 console RHIs)
  • Shading binning improvements optimized for console render backends (CMask/DCC, VRS)
    • This reduces bandwidth/tile overhead on platforms with CMask/DCC (typical of console GPUs).
  • Guidance for Nanite streaming pool sizing vs. GPU resource limits (console‑relevant)
    • The Cvar r.Nanite.Streaming.StreamingPoolSize now warns you to keep the pool size below typical 2–4 GB resource limits, which aligns with console buffer constraints
  • Optional skipping of Nanite shading dispatch while PSOs are still precaching (to avoid frame hitches on console)

If you have any specific questions, please don’t hesitate to ask, and I will do my best to answer them.

Cheers,

Tim

6

Hi, my reply late too! :slight_smile:

Thanks for the info. I guess the question was around any planned features for 5.8, 5.9 etc. If you knew there was some more perf to give but you’ve just not got round to it yet or that it’s running about as well as you intend.

But we’ll definitely give those 5.7 options a go when we switch.

Thanks

7

Hi again, I’m unable to provide specifics on performance improvements for 5.8 or 5.9 at this time, as we are still in the development cycle for 5.8, and any planning for 5.9 remains outstanding. However, we should have more concrete information closer to the actual release of 5.8. Try 5.7 to see how far it gets you, and if you encounter any pain points, please let us know. We can then look into getting you some targeted answers. I hope that works for you :slight_smile: