# r.SceneCulling.HierarchicalCPUCulling 

**URL:** <https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843>\
**Category:** Rendering\
**Tags:** unreal-engine, Epic-Pro-Support, question\
**Created:** [September 22, 2026, 9:32am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843 "2026-09-22T09:32:49Z")\
**Posts on this page:** 10\
**Page:** 1

<div class="post-metadata">

**Author:** ![anonymous-edc](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@anonymous-edc](https://forums.unrealengine.com/u/anonymous-edc)\
**Post date:** [September 22, 2026, 9:32am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/1 "2026-09-22T09:32:49Z")

</div>

Hi. I have a few questions on the new hierarchical cpu scene culling method you added in 5.8 (r.SceneCulling.HierarchicalCPUCulling).

1) I can only find 1 paragraph about this in the changelog. is there a documentation somewhere?

2) is it safe? how much is is tested? are you using this in production?

3) why nanite only? with a few mods it seems to work on non-nanite meshes too

4) ISM only, right? No HISM support

5) do you plan to add a PIE mode? an A/B test?

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![Ola.Olsson2](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@Ola.Olsson2](https://forums.unrealengine.com/u/Ola.Olsson2)\
**Post date:** [September 28, 2026, 7:30am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/2 "2026-09-28T07:30:52Z")

</div>

Hi.

1. No not yet.
2. It’s not very widely tested & considered experimental. I think it is used in production now or soon - but only in the context of PCG generated grass.
3. The CPU culling is only for NON-nanite so not sure what you are referring to here?
4. HISM has its own hierarchy and machinery and is something we’re slowly working on replacing (this is one step).
5. Yes!  
[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![anonymous-edc](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@anonymous-edc](https://forums.unrealengine.com/u/anonymous-edc)\
**Post date:** [September 28, 2026, 11:03am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/3 "2026-09-28T11:03:48Z")

</div>

3) I was probably wrong. I said that it’s nanite-only because from a fast look at the code it seems that the hierarchical data needs cooked data and it is built only if DoesTargetPlatformSupportNanite(TargetPlatform) is true. So, not really nanite-only but maybe nanite-support-only.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![Ola.Olsson2](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@Ola.Olsson2](https://forums.unrealengine.com/u/Ola.Olsson2)\
**Post date:** [September 28, 2026, 11:32am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/4 "2026-09-28T11:32:50Z")

</div>

Ah, right.

You are right, the hierarchy data is only produced on Nanite platforms at the moment and that is because before the CPU-side culling the only user is the GPU instance hierarchy. Now, it may make sense to do so but this has not really been tested yet. The runtime side excludes Nanite primitives (since they are GPU-culled), which is what I thought you were referring to. So, yes, it is indeed a small change to cook the data for all platforms.

I’d be curious to hear your experience with this if you try it more, since we don’t have a lot of internal data yet.

There’s been a few changes/fixes since 5.8 - notably if you have greatly varying instance scale it will not select LODs conservatively leading to missing data (this went in a week or so ago). But I’m also not 100% sure what state exactly it shipped in 5.8 TBH.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![Ola.Olsson2](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@Ola.Olsson2](https://forums.unrealengine.com/u/Ola.Olsson2)\
**Post date:** [September 29, 2026, 7:13am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/5 "2026-09-29T07:13:59Z")

</div>

I’m curious what problem you are wanting to solve with this? And also what platforms are you targeting for this?

The feature was added mainly for mobile because the mobile GPU-Scene implementation has to copy a lot of data for each instance and the conservative allocation sizes blow up. For desktop platforms it’s just an int32 per instance so generally ok to throw at the GPU. Except, also local lights can be pretty awful when just brute-forcing instances, so there’s that angle too.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![Ola.Olsson2](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@Ola.Olsson2](https://forums.unrealengine.com/u/Ola.Olsson2)\
**Post date:** [October 2, 2026, 10:39am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/6 "2026-10-02T10:39:58Z")

</div>

Those align with where we are going with this too which is good (EOL HISM).

I’m curious about specifics, like

\* the RVT speedups you mention - what situation is this important in? Do you have perf deltas that you could share?

\* any other cases that made you reach for the CPU-side culling?

Concrete data could be useful to prioritize this work internally.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![Ola.Olsson2](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@Ola.Olsson2](https://forums.unrealengine.com/u/Ola.Olsson2)\
**Post date:** [October 12, 2026, 7:37am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/7 "2026-10-12T07:37:47Z")

</div>

Thanks for the info!

For the legacy texture streaming part I think we tend to recommend using virtual textures since that solves this problem and more.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![anonymous-edc](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@anonymous-edc](https://forums.unrealengine.com/u/anonymous-edc)\
**Post date:** [September 29, 2026, 7:01am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/8 "2026-09-29T07:01:04Z")

</div>

I made various (but not so deep) changes:

1) to let this work on non nanite meshes too

2) to properly show the culled instance count

3) to hierarchical cull instances during rvt writing

this seems to work, looking at viewport and instance count. I didn’t do any gpu capture, hoping this is a mature feature missing only some minor changes.

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![anonymous-edc](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@anonymous-edc](https://forums.unrealengine.com/u/anonymous-edc)\
**Post date:** [October 2, 2026, 9:33am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/9 "2026-10-02T09:33:17Z")

</div>

We would like to use ISM for Nanite and non-nanite platforms too

We would like to remove HISMs due to their issues (per instance culling and lodding, cvar octree tuning, dynamic relevance in shadow)

We would like to optimize RVT rendering adding per instance culling

[Attachment Removed]

---

<div class="post-metadata">

**Author:** ![anonymous-edc](https://d3kjluh73b9h9o.cloudfront.net/original/4X/2/8/5/28502927f1d97ba1051b3b47ccfe17233bf41ef0.png) [@anonymous-edc](https://forums.unrealengine.com/u/anonymous-edc)\
**Post date:** [October 5, 2026, 8:51am UTC](https://forums.unrealengine.com/t/r-sceneculling-hierarchicalcpuculling/2853843/10 "2026-10-05T08:51:51Z")

</div>

We had an HISM with 20K instances mesh: it was ok to render that in the main camera (even without nanite). When we tried to render those only to RVT (since they were similar to mesh decals), fps collapsed. Each RVT tile was rendering all those 20K instances since the AABB was big as the world. So we has to split the HISM in smaller HISMs, so that the rvt culling could skip more HISMs AABBs. I don’t recall the exact timings, but we went from unshippable to 1 or 2 msecs tops.

Another use case may be CSM culling. We expect low cpu time when camera and sun don’t move. Now GetDynamicMeshElements tanks cpu when HISM are CastShadow. We had to implement a castshadow foliage splitter, to split the foliage hism in a low-instance-count + nearcamera castshadow foliage hism and a high-instance-count + farcamera no-castshadow hism.

Legacy streaming may use that system too. ISM takes the aabb for mipmap selection, forcing highest mipmap even for very far elements (consider 4 instances at the 4 corners of the world, with a camera at the origin). Considering each instance AABB may be overkill. Considering hierarchical aabb cells may be probably manageable.

I think that a generic API would be great for game developers, so that we can do volume queries more efficiently

[Attachment Removed]
