Wenbo Zhu
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Wenbo Zhu has a total of 20 co-authors. Here are the top co-authors by number of shared publications:
- Boyang (Tony) Yu
Georgia Institute of Technology, Atlanta, GA, USA
1 shared publication - Chaojian Li
Georgia Institute of Technology, Atlanta, GA, USA
1 shared publication - Cheng Wan
Georgia Institute of Technology, Atlanta, GA, USA
1 shared publication - Chenjie Weng
University of Southern California, Los Angeles, California, USA
1 shared publication - Christopher Batten
Cornell University, Ithaca, NY, USA
1 shared publication - Christopher Torng
University of Southern California, Los Angeles, California, USA
1 shared publication
Latest Publications
- Open Access
- 5 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-7
PowerFlow-DNN: Compiler-Directed Fine-Grained Power Orchestration for End-to-End Edge AI Inference
Edge AI systems operate under stringent energy and volume constraints, demanding extreme efficiency on limited battery capacity, with requirements worsening as intelligent capabilities advance. Prior work suggests fine-grained power orchestration through DVFS and power gating significantly improves efficiency critical to meeting such constraints, but introduces new challenges. We observe that layer-level approaches incur unintended overheads due to inter-layer coupling of power-control decisions, and jointly managing these mechanisms under limited voltage rails and transition overheads leads to a rapidly growing combinatorial schedule space. We propose PowerFlow-DNN, a compiler-directed framework for end-to-end power-state orchestration in ultra-low-power accelerators. By constructing a rigorous problem formulation for deadline-constrained, real-time, periodic inference as a unified inter-layer power-scheduling problem, our framework discovers energy-minimal power-state schedules while accounting for inter-layer impacts. We evaluate the framework on a DNN accelerator VLSI implementation in TSMC 40nm technology. Across representative edge networks, our approach discovers near-optimal solutions and achieves energy within 0.04% of the exact ILP oracle, reducing energy by up to 48% compared to an aggressive baseline without power orchestration, while reasoning over a combinatorial schedule space of over 10160 possible power-state assignments, yet operating on a structured layered state graph that enables efficient optimization, achieving up to 2.14× solver speedup via lightweight pruning.
- 5 August 2026
- Published by Association for Computing Machinery (ACM)
- p. 1-7
Open AccessPowerFlow-DNN: Compiler-Directed Fine-Grained Power Orchestration for End-to-End Edge AI Inference
Edge AI systems operate under stringent energy and volume constraints, demanding extreme efficiency on limited battery capacity, with requirements worsening as intelligent capabilities advance. Prior work suggests fine-grained power orchestration through DVFS and power gating significantly improves efficiency critical to meeting such constraints, but introduces new challenges. We observe that layer-level approaches incur unintended overheads due to inter-layer coupling of power-control decisions, and jointly managing these mechanisms under limited voltage rails and transition overheads leads to a rapidly growing combinatorial schedule space. We propose PowerFlow-DNN, a compiler-directed framework for end-to-end power-state orchestration in ultra-low-power accelerators. By constructing a rigorous problem formulation for deadline-constrained, real-time, periodic inference as a unified inter-layer power-scheduling problem, our framework discovers energy-minimal power-state schedules while accounting for inter-layer impacts. We evaluate the framework on a DNN accelerator VLSI implementation in TSMC 40nm technology. Across representative edge networks, our approach discovers near-optimal solutions and achieves energy within 0.04% of the exact ILP oracle, reducing energy by up to 48% compared to an aggressive baseline without power orchestration, while reasoning over a combinatorial schedule space of over 10160 possible power-state assignments, yet operating on a structured layered state graph that enables efficient optimization, achieving up to 2.14× solver speedup via lightweight pruning.
- Open Access
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1011-1025
Characterizing and Optimizing Realistic Workloads on a Commercial Compute-in-SRAM Device
- 17 October 2025
- Published by Association for Computing Machinery (ACM)
- p. 1011-1025
Open AccessCharacterizing and Optimizing Realistic Workloads on a Commercial Compute-in-SRAM Device
- Open Access
- 17 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Instant-3D: Instant Neural Radiance Field Training Towards On-Device AR/VR 3D Reconstruction
Neural Radiance Field (NeRF) based 3D reconstruction is highly desirable for immersive Augmented and Virtual Reality (AR/VR) applications, but achieving instant (i.e., < 5 seconds) on-device NeRF training remains a challenge. In this work, we first identify the inefficiency bottleneck: the need to interpolate NeRF embeddings up to 200,000 times from a 3D embedding grid during each training iteration. To alleviate this, we propose Instant-3D, an algorithm-hardware co-design acceleration framework that achieves instant on-device NeRF training. Our algorithm decomposes the embedding grid representation in terms of color and density, enabling computational redundancy to be squeezed out by adopting different (1) grid sizes and (2) update frequencies for the color and density branches. Our hardware accelerator further reduces the dominant memory accesses for embedding grid interpolation by (1) mapping multiple nearby points' memory read requests into one during the feed-forward process, (2) merging embedding grid updates from the same sliding time window during back-propagation, and (3) fusing different computation cores to support the different grid sizes needed by the color and density branches of Instant-3D algorithm. Extensive experiments validate the effectiveness of Instant-3D, achieving a large training time reduction of 41× - 248× while maintaining the same reconstruction quality. Excitingly, Instant-3D has enabled instant 3D reconstruction for AR/VR, requiring a reconstruction time of only 1.6 seconds per scene and meeting the AR/VR power consumption constraint of 1.9 W.
- 17 June 2023
- Published by Association for Computing Machinery (ACM)
- p. 1-13
Open AccessInstant-3D: Instant Neural Radiance Field Training Towards On-Device AR/VR 3D Reconstruction
Neural Radiance Field (NeRF) based 3D reconstruction is highly desirable for immersive Augmented and Virtual Reality (AR/VR) applications, but achieving instant (i.e., < 5 seconds) on-device NeRF training remains a challenge. In this work, we first identify the inefficiency bottleneck: the need to interpolate NeRF embeddings up to 200,000 times from a 3D embedding grid during each training iteration. To alleviate this, we propose Instant-3D, an algorithm-hardware co-design acceleration framework that achieves instant on-device NeRF training. Our algorithm decomposes the embedding grid representation in terms of color and density, enabling computational redundancy to be squeezed out by adopting different (1) grid sizes and (2) update frequencies for the color and density branches. Our hardware accelerator further reduces the dominant memory accesses for embedding grid interpolation by (1) mapping multiple nearby points' memory read requests into one during the feed-forward process, (2) merging embedding grid updates from the same sliding time window during back-propagation, and (3) fusing different computation cores to support the different grid sizes needed by the color and density branches of Instant-3D algorithm. Extensive experiments validate the effectiveness of Instant-3D, achieving a large training time reduction of 41× - 248× while maintaining the same reconstruction quality. Excitingly, Instant-3D has enabled instant 3D reconstruction for AR/VR, requiring a reconstruction time of only 1.6 seconds per scene and meeting the AR/VR power consumption constraint of 1.9 W.