Performance Implications of Multi-Chiplet Neural Processing Units on Autonomous Driving Perception
Not on the map. Why?
Abstract (arXiv)
We study the application of emerging chiplet-based Neural Processing Units to accelerate vehicular AI perception workloads in constrained automotive settings. The motivation stems from how chiplets technology is becoming integral to emerging vehicular architectures, providing a cost-effective trade-off between performance, modularity, and customization; and from perception models being the most computationally demanding workloads in a autonomous driving system. Using the Tesla Autopilot perception pipeline as a case study, we first breakdown its constituent models and profile their performance on different chiplet accelerators. From the insights, we propose a novel scheduling strategy to efficiently deploy perception workloads on multi-chip AI accelerators. Our experiments using a standard DNN performance simulator, MAESTRO, show our approach realizes 82% and 2.8x increase in throughput and processing engines utilization compared to monolithic accelerator designs.
Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Builds on 12 of its 41 references that are in the map. First posted on arXiv 24 Nov 2024.
Builds on
12 of 41 references.
References · 12
- 2017Attention Is All You Needcited by 59,932
- 2019EfficientDet: Scalable and Efficient Object Detectioncited by 1,018
- 2020Centaur: A Chiplet-based, Hybrid Sparse-Dense Accelerator for Personalized Recommendationscited by 21
- 2022HydraFusion: Context-Aware Selective Sensor Fusion for Robust and Efficient Autonomous Vehicle Perceptioncited by 16
- 2023Map-and-Conquer: Energy-Efficient Mapping of Dynamic Neural Nets onto Heterogeneous MPSoCscited by 9
- 2021SAGE: A Split-Architecture Methodology for Efficient End-to-End Autonomous Vehicle Controlcited by 8
- 2022DREAM: A Dynamic Scheduler for Dynamic Real-time Multi-model ML Workloadscited by 8
- 2023Gemini: Mapping and Architecture Co-exploration for Large-scale DNN Chiplet Acceleratorscited by 7
- 2022H2H: Heterogeneous Model to Heterogeneous System Mapping with Computation and Communication Awarenesscited by 6
- 2024SCAR: Scheduling Multi-Model AI Workloads on Heterogeneous Multi-Chiplet Module Acceleratorscited by 6
- 2022Romanus: Robust Task Offloading in Modular Multi-Sensor Autonomous Driving Systemscited by 3
- 2024CARIn: Constraint-Aware and Responsive Inference on Heterogeneous Devices for Single- and Multi-DNN Workloadscited by 1
Built on by
Citing papers per year: 2026: 1.
1 paper.