Skip to content
Preprint

GSPIM: A Temporal-Locality-Aware 4D Gaussian Splatting Accelerator with Polymorphic PIM and Multi-Frame Rendering Dataflow

Nov 2026 · 0 citations · 68 references
Computer Science

TL;DR

GSPIM is presented, a temporal-locality-aware 4DGS accelerator for PIM-enabled edge devices that achieves a 4.15x geometric-mean speedup and a 4.99x geometric-mean energy-efficiency improvement over a fully optimized frame-by-frame GPU baseline.

Abstract

4D Gaussian Splatting (4DGS) enables high-quality real-time novel-view synthesis for dynamic scenes, supporting free-viewpoint video, mobile XR, and autonomous systems at the edge. Methods using explicit 4D distributions, keyframe interpolation, or anchor binding share two GPU data-movement bottlenecks. Temporal filtering follows model record loading, wasting LPDDR5 bandwidth on inactive data. Frame-isolated execution also overlooks highly overlapping active sets and depth relationships across adjacent frames, reloading shared attributes when working sets exceed on-chip cache capacity. To address these bottlenecks, we present GSPIM, a temporal-locality-aware 4DGS accelerator for PIM-enabled edge devices. Polymorphic PIM adapts a shared near-memory datapath to representation-specific field layouts and selection procedures. It filters and compacts valid data into contiguous inputs before the GPU accesses complete attributes, reducing ineffective transfers. Multi-Frame Rendering Dataflow reuses active attributes and stable depth relationships within a temporal window, reducing repeated work across frames. Hybrid Dataflow Scheduling overlaps near-memory data preparation with GPU rendering to reduce waiting. Evaluation combines a GPU-PIM model calibrated to Jetson AGX Orin 64 GB with a 1z DRAM physical implementation of the added circuits. Across three 4DGS methods and six dynamic scenes, GSPIM achieves a 4.15x geometric-mean speedup and a 4.99x geometric-mean energy-efficiency improvement over a fully optimized frame-by-frame GPU baseline. The maximum scene-level PSNR difference is 0.011 dB, and the added physical footprint is 18.7172 square millimeters across 32 LPDDR5 dies.

View source

Similar papers

Book Open access 2027

SL²OT: Scaling LLM Locality on Tile with Distributed 3D-DRAM as Unified Memory

Large language models (LLMs) are increasingly limited by memory bandwidth and latency rather than compute throughput. Our modeled 3D-DRAM system offers 4× the raw bandwidth of the HBM reference. However, high remote-access latency (~3× local latency) and redundant NoC traffic reduce its effective bandwidth below that o...

Hao-Jia Hui, Mo Guang, Ya-Qian Li et al. · 0 citations
Preprint Oct 2026

ByteSplat: Efficient Distributed 3D Gaussian Splatting Training via Intra- and Inter-GPU communication reduction

3D Gaussian Splatting (3DGS) enables photorealistic scene reconstruction, but training large-scale scenes requires substantial memory and computation. Distributing training across multiple GPUs increases available memory capacity, yet its performance is strictly constrained by data movement. We identify two dominant bo...

Shuo Wu, He Zhu, Han Zhao et al. · 0 citations
Preprint Oct 2026

M3D-FLOW: A Monolithic 3-D-Stackable Processing-Near-DRAM Architecture for Generative Normalizing Flow Applications

A novel monolithic 3D-stackable (M3D) DRAM-based processing-near-memory (PNM) architecture that eliminates memory bottlenecks using a Cu-Cu hybrid-bonded logic die integrated with an M3D DRAM die is proposed, and hybrid binary--stochastic floating-point multipliers are employed.

Bipin Thapa Magar, I. Thakkar, Sudeep Pasricha · 0 citations
Preprint Oct 2026

Understanding NVIDIA Grace-Blackwell Unified Memory Through Placement-Controlled Microbenchmarks

NVIDIA Grace--Blackwell systems make CPU--GPU memory increasingly unified at the programming level, but unified addressability does not make the memory system uniform. A CUDA managed allocation on GB200 may be served from GPU-local HBM, accessed through host/system memory, demand-migrated, prefetched, reused after GPU...

Nathan Graddon, Aaron Jarmusch, Sunita Chandrasekaran · 0 citations
Preprint Oct 2026

Compiling Semi-Ring Dynamic Programming to Tier-Aware 3D-DRAM Processing-in-Memory

Processing-in-Memory (PIM) on monolithic 3D (M3D) DRAM is a promising answer to the memory wall for data-intensive dynamic programming (DP), yet extracting its performance today demands hand-written kernels: the programmer must pick a tile size, place data across non-uniform-latency memory tiers, partition work between...

Mahbod Afarin, Tsung-Han Lu, T. Rosing · 0 citations

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.