Metraipolis › archive › dc_0628cbbcc24d

GSPIM: A Temporal-Locality-Aware 4DGS Accelerator for PIM-Enabled Edge Devices — arXiv:2611.04804

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.04804
License (as archived)
CC0-1.0
Archived
2026-11-04T21:46:11.804Z
SHA-256 of the archived text
cb1cc6657cb3b539068f827683ed87a74f0a15dfbb963fde79a6997a8062d9ae
Stable URL of this page
https://metraipolis.com/documento/dc_0628cbbcc24d

How to cite

GSPIM: A Temporal-Locality-Aware 4DGS Accelerator for PIM-Enabled Edge Devices — arXiv:2611.04804 — arXiv (en), https://arxiv.org/abs/2611.04804 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_0628cbbcc24d (Metraipolis archive dc_0628cbbcc24d, sha256 cb1cc6657cb3b539068f827683ed87a74f0a15dfbb963fde79a6997a8062d9ae)
BibTeX
@misc{dc0628cbbcc24d,
  title = {GSPIM: A Temporal-Locality-Aware 4DGS Accelerator for PIM-Enabled Edge Devices — arXiv:2611.04804},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.04804},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_0628cbbcc24d; sha256 cb1cc6657cb3b539068f827683ed87a74f0a15dfbb963fde79a6997a8062d9ae},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.04804. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

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.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.AR, nuovi lavori del giorno).
Categorie arXiv: Hardware Architecture (cs.AR).
Pagina del lavoro: https://arxiv.org/abs/2611.04804 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-04T21:36:34.396Z dalla pagina https://arxiv.org/list/cs.AR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).