pith. sign in

arxiv: 2606.11529 · v1 · pith:TKZWUL6Tnew · submitted 2026-06-10 · 💻 cs.GR · cs.CV· cs.PF

XPR: An Extensible Cross-Platform Point-Based Differentiable Renderer

Steve Rhyner , Sankeerth Durvasula , Aleksandr Kovalev , Hansel Jia , Adrian Zhao , Mrutunjayya Mrutunjayya , Nilesh Ahuja , Selvakumar Panneer
show 2 more authors
Christina Giannoula Nandita Vijaykumar
This is my paper

Pith reviewed 2026-06-27 07:57 UTC · model grok-4.3

classification 💻 cs.GR cs.CVcs.PF
keywords point-based renderingdifferentiable renderingcross-platformXLA compilation3D Gaussian splattingnovel view synthesis3D reconstruction
0
0 comments X

The pith

XPR lets new point-based renderers be written in a few hundred lines of Python and compiled across GPUs, TPUs and CPUs via XLA.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

Point-based differentiable rendering supports 3D reconstruction and novel-view synthesis, yet new methods usually demand extensive low-level kernels and manual gradient code that tie them to specific hardware. XPR supplies a high-level interface that isolates method-specific logic from a shared pipeline. The pipeline is broken into modular, statically shaped parallel operations that an XLA compiler can lower to multiple accelerator types. The paper shows that 3DGS, 3DGUT and LinPrim can each be expressed in only a few hundred lines of Python under this interface. The result is portable execution without per-platform rewrites or hand-written backward passes.

Core claim

XPR decomposes point-based differentiable rendering into modular, statically shaped parallel operations that separate method-specific logic from the shared pipeline. These operations are lowered by the XLA compiler to GPUs, TPUs, CPUs and other ML accelerators. Under this structure, complete implementations of 3DGS, 3DGUT and LinPrim are each expressed in a few hundred lines of Python and execute portably without hardware-specific kernels or manually written backward passes.

What carries the argument

The decomposition of rendering into modular, statically shaped parallel operations that isolate method-specific logic from the shared pipeline.

If this is right

  • New point-based methods can be prototyped without writing hardware-specific kernels.
  • Backward passes are obtained automatically from the modular forward operations.
  • The same Python source runs on GPUs, TPUs, CPUs and other XLA-supported accelerators.
  • Reproducibility improves because method logic is expressed at a high level rather than in low-level code.
  • Emerging rendering systems become deployable across diverse hardware without repeated porting effort.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same separation of logic from pipeline could be tested on non-point-based differentiable graphics pipelines.
  • If the modular operations scale, the approach might support larger scene representations without custom memory management.
  • Portability to new accelerators would let researchers move experiments from research clusters to edge devices with minimal code changes.

Load-bearing premise

The modular decomposition into statically shaped parallel operations preserves both performance and correct differentiability once lowered by the XLA compiler.

What would settle it

An XPR implementation of 3DGS that produces measurably different images or incorrect gradients compared with a reference hand-written kernel on the same input and hardware would falsify the claim.

Figures

Figures reproduced from arXiv: 2606.11529 by Adrian Zhao, Aleksandr Kovalev, Christina Giannoula, Hansel Jia, Mrutunjayya Mrutunjayya, Nandita Vijaykumar, Nilesh Ahuja, Sankeerth Durvasula, Selvakumar Panneer, Steve Rhyner.

Figure 1
Figure 1. Figure 1: Projection and per-pixel evaluation for three representative [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 3
Figure 3. Figure 3: Parallelizing the preprocess and filter stages using [PITH_FULL_IMAGE:figures/full_fig_p004_3.png] view at source ↗
Figure 4
Figure 4. Figure 4: Modules used to parallelize the rasterizer. [PITH_FULL_IMAGE:figures/full_fig_p005_4.png] view at source ↗
Figure 5
Figure 5. Figure 5: Parallelizing the shader using modules. 4 Detailed Design 4.1 Preprocess and Filter Preprocess. XPR expresses this stage as map_fn that uses project and 𝑁 primitives as arguments to expose parallelism. Each primi￾tive is processed independently by hardware processing units. The operation project computes the projected footprint, bounding box, depth, a visibility flag, shader data, and tile cull data per pr… view at source ↗
Figure 6
Figure 6. Figure 6: XPR’s rasterizer using modules over static-shape arrays. 5 [PITH_FULL_IMAGE:figures/full_fig_p005_6.png] view at source ↗
Figure 7
Figure 7. Figure 7: The shader receives 𝐿 bins. Each bin contains tiles whose primitive counts are bounded by one static trip count. The shader processes each tile’s primitive list in batches of 𝑏𝑙 entries. Algorithm 1 describes the process for a bin 𝑙. For each tile 𝑡, the shader receives as input the trip count 𝑃𝑙 , the batch size 𝑏𝑙 , and the list 𝑄𝑡 of primitive indices to iterate through. The shader loops across 𝑄𝑡 primi… view at source ↗
Figure 8
Figure 8. Figure 8: Average rendering FPS (y-axis) vs. the number of bins (x-axis) [PITH_FULL_IMAGE:figures/full_fig_p013_8.png] view at source ↗
read the original abstract

Point-based differentiable rendering underpins modern 3D reconstruction, novel-view synthesis, and learning-based graphics pipelines, but developing new rendering methods often requires extensive low-level implementation, hardware-specific kernels, and manually written backward passes. This limits rapid prototyping, reproducibility, exploration, and deployment, especially across diverse hardware platforms. This paper presents XPR, an extensible cross-platform framework for point-based differentiable rendering. XPR introduces a high-level programming interface that separates method-specific logic from the shared rendering pipeline, allowing users to implement new methods in a few lines of code. Its pipeline decomposes rendering into modular, statically shaped parallel operations that can be lowered by a cross-platform compiler to GPUs, TPUs, CPUs, and other ML accelerators. We demonstrate implementations of 3DGS, 3DGUT, and LinPrim, with only a few 100s lines of Python code, each of which can be compiled to a range of hardware platforms with the XLA compiler. These results show that XPR enables fast experimentation and portable execution for emerging point-based differentiable rendering systems.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 0 minor

Summary. The paper presents XPR, an extensible cross-platform framework for point-based differentiable rendering. It introduces a high-level programming interface that separates method-specific logic from a shared rendering pipeline decomposed into modular, statically shaped parallel operations. These operations are designed to be lowered by the XLA compiler to GPUs, TPUs, CPUs, and other ML accelerators. The paper demonstrates the approach by providing implementations of 3DGS, 3DGUT, and LinPrim, each in only a few hundred lines of Python code.

Significance. If the central claims hold, XPR would meaningfully lower the barrier to developing, reproducing, and deploying new point-based differentiable renderers by reducing the need for hand-written kernels and backward passes. The emphasis on a reusable, compiler-lowered pipeline and cross-platform portability via XLA addresses a practical pain point in the field. The provision of concise Python implementations for established methods is a concrete strength that could aid reproducibility if the code is released.

major comments (2)
  1. [Abstract] Abstract: the assertion that the XLA-lowered implementations achieve practical performance and correct differentiability for 3DGS, 3DGUT, and LinPrim is load-bearing for the central claim, yet the manuscript supplies no runtime measurements, memory usage figures, gradient verification against reference implementations, or hardware-specific results.
  2. [Pipeline] Shared pipeline description: the decomposition into statically shaped parallel operations must handle data-dependent steps such as per-point depth sorting, covariance projection, and alpha blending; without explicit treatment of how variable cardinality is managed (e.g., via padding or masking) it is unclear whether exact gradient flow and competitive runtime are preserved when XLA lowers the kernels.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback highlighting the need for quantitative evidence and explicit pipeline details. We address each major comment below and will revise the manuscript accordingly to strengthen the central claims.

read point-by-point responses

  1. Referee: [Abstract] Abstract: the assertion that the XLA-lowered implementations achieve practical performance and correct differentiability for 3DGS, 3DGUT, and LinPrim is load-bearing for the central claim, yet the manuscript supplies no runtime measurements, memory usage figures, gradient verification against reference implementations, or hardware-specific results.

    Authors: We agree that the manuscript currently lacks quantitative runtime measurements, memory usage figures, gradient verification, and hardware-specific results, which are needed to fully support the claims of practical performance and correct differentiability. The existing demonstrations emphasize implementation conciseness and XLA compilability rather than benchmarks. In the revision we will add these evaluations, including timing and memory comparisons against reference implementations on multiple platforms, plus numerical gradient checks. revision: yes

  2. Referee: [Pipeline] Shared pipeline description: the decomposition into statically shaped parallel operations must handle data-dependent steps such as per-point depth sorting, covariance projection, and alpha blending; without explicit treatment of how variable cardinality is managed (e.g., via padding or masking) it is unclear whether exact gradient flow and competitive runtime are preserved when XLA lowers the kernels.

    Authors: The design relies on statically shaped operations for XLA compatibility, with variable point counts managed via padding to a fixed maximum cardinality and element-wise masking. Masking ensures padded elements contribute neither to forward passes (e.g., sorting, projection, blending) nor to gradients, preserving exact differentiability. We will add an explicit subsection describing this padding/masking mechanism, its impact on gradient flow, and why it maintains competitive runtime under XLA lowering. revision: yes

Circularity Check

0 steps flagged

No circularity: framework presentation with no derived predictions or fitted results

full rationale

The paper presents a software framework (XPR) for point-based differentiable rendering, with claims centered on code brevity for implementing methods like 3DGS and compilation via XLA to multiple platforms. No equations, first-principles derivations, parameter fitting, or predictions appear in the provided text. Claims are demonstrated through implementation examples rather than any reduction to prior inputs or self-citations. This is a systems contribution evaluated by engineering demonstration, not a mathematical result that could be circular by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests on the capability of an existing compiler (XLA) to handle the modular operations and on the assumption that high-level Python descriptions can be lowered without breaking differentiability or performance; no free parameters or new entities are introduced.

axioms (1)
  • domain assumption The XLA compiler can lower the modular, statically shaped parallel operations to target hardware platforms while preserving differentiability.
    Invoked when the paper states that the pipeline can be compiled to GPUs, TPUs, CPUs and other ML accelerators.

pith-pipeline@v0.9.1-grok · 5771 in / 1160 out tokens · 18622 ms · 2026-06-27T07:57:47.403942+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

63 extracted references · 3 canonical work pages · 2 internal anchors

  1. [1]

    Jason Ansel, Edward Yang, Horace He, Natalia Gimelshein, Animesh Jain, Michael Voznesensky, Bin Bao, Peter Bell, David Berard, Evgeni Burovski, et al . 2024. PyTorch 2: Faster Machine Learning Through Dynamic Python Bytecode Trans- formation and Graph Compilation. InProceedings of the 29th ACM international conference on architectural support for programm...

    2024

  2. [2]

    AWS. 2026. Neuron Graph Compiler. AWS Neuron Documentation. https: //awsdocs-neuron.readthedocs-hosted.com/en/latest/compiler/index.html ac- cessed: May 7, 2026

    2026

  3. [3]

    AWS. 2026. PyTorch NeuronX Tracing API for Inference. AWS Neuron Documentation. https://awsdocs-neuron.readthedocs-hosted.com/en/latest/ frameworks/torch/torch-neuronx/api-reference-guide/inference/api-torch- neuronx-trace.html accessed: May 7, 2026

    2026

  4. [4]

    Jonathan T Barron, Ben Mildenhall, Dor Verbin, Pratul P Srinivasan, and Pe- ter Hedman. 2022. Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 5470–5479

    2022

  5. [5]

    Oliver Batchelor. 2023. Taichi Splatting. https://github.com/uc-vision/taichi- splatting. accessed: May 7, 2026

    2023

  6. [6]

    2018.JAX: composable transformations of Python+NumPy programs

    James Bradbury, Roy Frostig, Peter Hawkins, Matthew James Johnson, Yash Katariya, Chris Leary, Dougal Maclaurin, George Necula, Adam Paszke, Jake VanderPlas, Skye Wanderman-Milne, and Qiao Zhang. 2018.JAX: composable transformations of Python+NumPy programs. http://github.com/jax-ml/jax

    2018

  7. [7]

    Arthur Brussee. 2024. Brush: 3D Reconstruction for All. https://github.com/ ArthurBrussee/brush accessed: May 7, 2026

    2024

  8. [8]

    Yuanhao Cai, Yixun Liang, Jiahao Wang, Angtian Wang, Yulun Zhang, Xiaokang Yang, Zongwei Zhou, and Alan Yuille. 2024. Radiative Gaussian Splatting for Efficient X-Ray Novel View Synthesis. InEuropean Conference on Computer Vision. Springer, 283–299

    2024

  9. [9]

    Pinxuan Dai, Jiamin Xu, Wenxiang Xie, Xinguo Liu, Huamin Wang, and Weiwei Xu. 2024. High-quality Surface Reconstruction using Gaussian Surfels. InACM SIGGRAPH 2024 conference papers. 1–11

    2024

  10. [10]

    Sankeerth Durvasula, Sharanshangar Muhunthan, Zain Moustafa, Richard Chen, Ruofan Liang, Yushi Guan, Nilesh Ahuja, Nilesh Jain, Selvakumar Panneer, and Nandita Vijaykumar. 2025. ContraGS: Codebook-Condensed and Trainable Gauss- ian Splatting for Fast, Memory-Efficient Reconstruction. InProceedings of the IEEE/CVF International Conference on Computer Visio...

    2025

  11. [11]

    Guofeng Feng, Siyan Chen, Rong Fu, Zimu Liao, Yi Wang, Tao Liu, Boni Hu, Linning Xu, Zhilin Pei, Hengjie Li, et al. 2025. FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering. InProceedings of the Computer Vision and Pattern Recognition Conference. 26652–26662

    2025

  12. [12]

    2024.Kaolin: A PyTorch Library for Accelerating 3D Deep Learning Research

    Clement Fuji Tsang, Maria Shugrina, Jean Francois Lafleche, Or Perel, Charles Loop, Towaki Takikawa, Vismay Modi, Alexander Zook, Jiehan Wang, Wenzheng Chen, Tianchang Shen, Jun Gao, Krishna Murthy Jatavallabhula, Edward Smith, Artem Rozantsev, Sanja Fidler, Gavriel State, Jason Gorski, Tommy Xiang, Jianing Li, Michael Li, and Rev Lebaredian. 2024.Kaolin:...

    2024

  13. [13]

    Ruiqi Gao, Aleksander Holynski, Philipp Henzler, Arthur Brussee, Ricardo Mar- tin Brualla, Pratul Srinivasan, Jonathan Barron, and Ben Poole. 2024. CAT3D: Create Anything in 3D with Multi-View Diffusion Models.Advances in Neural Information Processing Systems37 (2024), 75468–75494

    2024

  14. [14]

    Fredrik Gustafsson and Gustaf Hendeby. 2011. Some Relations Between Extended and Unscented Kalman Filters.IEEE Transactions on Signal Processing60, 2 (2011), 545–555

    2011

  15. [15]

    Abdullah Hamdi, Luke Melas-Kyriazi, Jinjie Mai, Guocheng Qian, Ruoshi Liu, Carl Vondrick, Bernard Ghanem, and Andrea Vedaldi. 2024. GES : Generalized Exponential Splatting for Efficient Radiance Field Rendering. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 19812–19822

    2024

  16. [16]

    Jan Held, Renaud Vandeghen, Adrien Deliege, Abdullah Hamdi, Daniel Rebain, Silvio Giancola, Anthony Cioppa, Andrea Vedaldi, Bernard Ghanem, Andrea Tagliasacchi, et al. 2026. Triangle Splatting for Real-Time Radiance Field Render- ing. InThirteenth International Conference on 3D Vision

    2026

  17. [17]

    Yuanming Hu, Tzu-Mao Li, Luke Anderson, Jonathan Ragan-Kelley, and Frédo Durand. 2019. Taichi: A Language for High-Performance Computation on Spa- tially Sparse Data Structures.ACM Transactions on Graphics (TOG)38, 6 (2019), 1–16

    2019

  18. [18]

    Binbin Huang, Zehao Yu, Anpei Chen, Andreas Geiger, and Shenghua Gao. 2024. 2D Gaussian Splatting for Geometrically Accurate Radiance Fields. InACM SIGGRAPH 2024 conference papers. 1–11

    2024

  19. [19]

    Tianyu Huang, Wangguandong Zheng, Tengfei Wang, Yuhao Liu, Zhenwei Wang, Junta Wu, Jie Jiang, Hui Li, Rynson Lau, Wangmeng Zuo, et al. 2025. Voyager: Long-Range and World-Consistent Video Diffusion for Explorable 3D Scene Generation.ACM Transactions on Graphics (TOG)44, 6 (2025), 1–15

    2025

  20. [20]

    2022.Mitsuba 3 renderer

    Wenzel Jakob, Sébastien Speierer, Nicolas Roussel, Merlin Nimier-David, Delio Vicini, Tizian Zeltner, Baptiste Nicolet, Miguel Crespo, Vincent Leroy, and Ziyi Zhang. 2022.Mitsuba 3 renderer. https://mitsuba-renderer.org

    2022

  21. [21]

    Norman P Jouppi, Cliff Young, Nishant Patil, David Patterson, Gaurav Agrawal, Raminder Bajwa, Sarah Bates, Suresh Bhatia, Nan Boden, Al Borchers, et al. 2017. In-Datacenter Performance Analysis of a Tensor Processing Unit. InProceedings of the 44th annual international symposium on computer architecture. 1–12

    2017

  22. [22]

    Nikhil Keetha, Jay Karhade, Krishna Murthy Jatavallabhula, Gengshan Yang, Se- bastian Scherer, Deva Ramanan, and Jonathon Luiten. 2024. SplaTAM: Splat Track & Map 3D Gaussians for Dense RGB-D SLAM. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 21357–21366

    2024

  23. [23]

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, and George Drettakis

  24. [24]

    https: //github.com/graphdeco-inria/gaussian-splatting

    3D Gaussian Splatting for Real-Time Radiance Field Rendering. https: //github.com/graphdeco-inria/gaussian-splatting. accessed: May 7, 2026

    2026

  25. [25]

    Bernhard Kerbl, Georgios Kopanas, Thomas Leimkühler, George Drettakis, et al

  26. [26]

    Graph.42, 4 (2023)

    3D Gaussian Splatting for Real-Time Radiance Field Rendering.ACM Trans. Graph.42, 4 (2023)

    2023

  27. [27]

    Arno Knapitsch, Jaesik Park, Qian-Yi Zhou, and Vladlen Koltun. 2017. Tanks and Temples: Benchmarking Large-Scale Scene Reconstruction.ACM Transactions on Graphics (ToG)36, 4 (2017), 1–13

    2017

  28. [28]

    George Kopanas. 2024. Slang.D Gaussian Splatting Rasterizer. https://github. com/google/slang-gaussian-rasterization. accessed: May 7, 2026

    2024

  29. [29]

    Samuli Laine, Janne Hellsten, Tero Karras, Yeongho Seol, Jaakko Lehtinen, and Timo Aila. 2020. Modular Primitives for High-Performance Differentiable Ren- dering.ACM Transactions on Graphics (ToG)39, 6 (2020), 1–14

    2020

  30. [30]

    Wanhua Li, Yujie Zhao, Minghan Qin, Yang Liu, Yuanhao Cai, Chuang Gan, and Hanspeter Pfister. 2025. LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS.Advances in Neural Information Processing Systems38 (2025)

    2025

  31. [31]

    Guanxing Lu, Shiyi Zhang, Ziwei Wang, Changliu Liu, Jiwen Lu, and Yansong Tang. 2024. ManiGaussian: Dynamic Gaussian Splatting for Multi-task Robotic Manipulation. InEuropean Conference on Computer Vision. Springer, 349–366

    2024

  32. [32]

    Daniel MacSwayne, Ales Leonardis, and Jianbo Jiao. 2026. 3D Superquadric Splatting. InProceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision. 5154–5163

    2026

  33. [33]

    Alexander Mai, Peter Hedman, George Kopanas, Dor Verbin, David Futschik, Qiangeng Xu, Falko Kuester, Jonathan T Barron, and Yinda Zhang. 2025. EVER: Exact Volumetric Ellipsoid Rendering for Real-time View Synthesis. InProceedings of the IEEE/CVF International Conference on Computer Vision. 4930–4939

    2025

  34. [34]

    Hidenobu Matsuki, Riku Murai, Paul HJ Kelly, and Andrew J Davison. 2024. Gaussian Splatting SLAM. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 18039–18048

    2024

  35. [35]

    Nelson Max. 1995. Optical Models for Direct Volume Rendering.IEEE Transactions on Visualization and Computer Graphics1, 2 (1995), 99–108

    1995

  36. [36]

    Ben Mildenhall, Pratul P Srinivasan, Matthew Tancik, Jonathan T Barron, Ravi Ramamoorthi, and Ren Ng. 2021. NeRF: representing scenes as neural radiance fields for view synthesis.Commun. ACM65, 1 (2021), 99–106

    2021

  37. [37]

    NVIDIA. 2025. Vulkan Gaussian Splatting. https://github.com/nvpro-samples/ vk_gaussian_splatting accessed: May 7, 2026

    2025

  38. [38]

    NVIDIA Spatial Intelligence Lab. 2025. Official Implementation of 3D Gaussian Unscented Transform. https://github.com/nv-tlabs/3dgrut. accessed: May 7, 2026

    2025

  39. [39]

    OpenXLA Project. 2024. StableHLO. OpenXLA Documentation. https://openxla. org/stablehlo accessed: May 7, 2026

    2024

  40. [40]

    OpenXLA Project. 2024. XLA: Accelerated Linear Algebra. https://openxla.org/ xla accessed: May 7, 2026

    2024

  41. [41]

    Jaesung Park. 2024. vkgs: Vulkan-based Gaussian Splatting Viewer. https: //github.com/jaesung-cs/vkgs

    2024

  42. [42]

    Jack Parker-Holder, Philip Ball, Jake Bruce, Vibhavari Dasagi, Kristian Hol- sheimer, Christos Kaplanis, Alexandre Moufarek, Guy Scully, Jeremy Shar, Jimmy Shi, Stephen Spencer, Jessica Yung, Michael Dennis, Sultan Kenjeyev, Shangbang Long, Vlad Mnih, Harris Chan, Maxime Gazeau, Bonnie Li, Fabio Pardo, Luyu Wang, Lei Zhang, Frederic Besse, Tim Harley, Ann...

    2024

  43. [43]

    Ben Poole, Ajay Jain, Jonathan T Barron, and Ben Mildenhall. 2022. DreamFusion: Text-to-3D using 2D Diffusion.arXiv preprint arXiv:2209.14988(2022)

    work page internal anchor Pith review Pith/arXiv arXiv 2022

  44. [44]

    Jiating Qian, Yiming Yan, Fengjiao Gao, Baoyu Ge, Maosheng Wei, Boyi Shang- guan, and Guangjun He. 2025. C3DGS: Compressing 3D Gaussian Model for Surface Reconstruction of Large-Scale Scenes Based on Multiview UAV Images. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 18 (2025), 4396–4409

    2025

  45. [45]

    Minghan Qin, Wanhua Li, Jiawei Zhou, Haoqian Wang, and Hanspeter Pfister

  46. [46]

    InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition

    LangSplat: 3D Language Gaussian Splatting. InProceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. 20051–20060

  47. [47]

    Lukas Radl, Michael Steiner, Mathias Parger, Alexander Weinrauch, Bernhard Kerbl, and Markus Steinberger. 2024. StopThePop: Sorted Gaussian Splatting for View-Consistent Real-time Rendering.ACM Transactions on Graphics (TOG)43, 4 (2024), 1–17. 9

    2024

  48. [48]

    Nikhila Ravi, Jeremy Reizenstein, David Novotny, Taylor Gordon, Wan-Yen Lo, Justin Johnson, and Georgia Gkioxari. 2020. Accelerating 3D Deep Learning with PyTorch3D.arXiv:2007.08501(2020)

    work page internal anchor Pith review Pith/arXiv arXiv 2020

  49. [49]

    James Reed, Zachary DeVito, Horace He, Ansley Ussery, and Jason Ansel. 2022. torch.fx: Practical Program Capture and Transformation for Deep Learning in Python.Proceedings of Machine Learning and Systems4 (2022), 638–651

    2022

  50. [50]

    shg8. 2024. 3DGS.cpp. https://github.com/shg8/3DGS.cpp accessed: May 7, 2026

    2024

  51. [51]

    Yusuke Takimoto, Hiroyuki Sato, Hikari Takehara, Keishiro Uragaki, Takehiro Tawara, Xiao Liang, Kentaro Oku, Wataru Kishimoto, and Bo Zheng. 2022. Dressi: A Hardware-Agnostic Differentiable Renderer with Reactive Shader Packing and Soft Rasterization. InComputer Graphics Forum, Vol. 41. Wiley Online Library, 13–27

    2022

  52. [52]

    Jiaxiang Tang, Jiawei Ren, Hang Zhou, Ziwei Liu, and Gang Zeng. 2024. Dream- Gaussian: Generative Gaussian Splatting for Efficient 3D Content Creation. In International Conference on Learning Representations

    2024

  53. [53]

    HunyuanWorld Team, Zhenwei Wang, Yuhao Liu, Junta Wu, Zixiao Gu, Haoyuan Wang, Xuhui Zuo, Tianyu Huang, Wenhuan Li, Sheng Zhang, et al. 2025. Hun- yuanWorld 1.0: Generating Immersive, Explorable, and Interactive 3D Worlds from Words or Pixels.arXiv preprint arXiv:2507.21809(2025)

    work page arXiv 2025

  54. [54]

    Piero Toffanin. 2024. OpenSplat. https://github.com/pierotofy/opensplat ac- cessed: May 7, 2026

    2024

  55. [55]

    Nicolas von Lützow and Matthias Nießner. 2025. LinPrim: Linear Primitives for Differentiable Volumetric Rendering. InThe Thirty-ninth Annual Conference on Neural Information Processing Systems

    2025

  56. [56]

    Nicolas von Lützow and Matthias Nießner. 2025. LinPrim: Linear Primitives for Differentiable Volumetric Rendering. https://github.com/nicolasvonluetzow/ linear-splatting. accessed: May 7, 2026

    2025

  57. [57]

    Qi Wu, Janick Martinez Esturo, Ashkan Mirzaei, Nicolas Moenne-Loccoz, and Zan Gojcic. 2025. 3DGUT: Enabling Distorted Cameras and Secondary Rays in Gaussian Splatting. InProceedings of the Computer Vision and Pattern Recognition Conference. 26036–26046

    2025

  58. [58]

    Chi Yan, Delin Qu, Dan Xu, Bin Zhao, Zhigang Wang, Dong Wang, and Xuelong Li

  59. [59]

    InProceedings of the IEEE/CVF conference on computer vision and pattern recognition

    GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 19595–19604

  60. [60]

    Vickie Ye, Ruilong Li, Justin Kerr, Matias Turkulainen, Brent Yi, Zhuoyang Pan, Otto Seiskari, Jianbo Ye, Jeffrey Hu, Matthew Tancik, et al. 2025. gsplat: An Open- Source Library for Gaussian Splatting.Journal of Machine Learning Research26, 34 (2025), 1–17

    2025

  61. [61]

    Zehao Yu, Anpei Chen, Binbin Huang, Torsten Sattler, and Andreas Geiger. 2024. Mip-Splatting: Alias-free 3D Gaussian Splatting. InProceedings of the IEEE/CVF conference on computer vision and pattern recognition. 19447–19456

    2024

  62. [62]

    Baowen Zhang, Chuan Fang, Rakesh Shrestha, Yixun Liang, Xiao-Xiao Long, and Ping Tan. 2026. RaDe-GS: Rasterizing Depth in Gaussian Splatting.ACM Transactions on Graphics45, 2 (2026), 1–14

    2026

  63. [63]

    Matthias Zwicker, Hanspeter Pfister, Jeroen Van Baar, and Markus Gross. 2002. EWA Splatting.IEEE Transactions on Visualization & Computer Graphics8, 03 (2002), 223–238. A Evaluation Details A.1 Ease of Implementing Rendering Methods We demonstrate that XPR can express various point-based ren- dering methods with only small code changes. XPR’s high-level p...

    2002