Published in ASPLOS'2024', 2024
The first system that enables distributed training on mobile-SoC-Clustered edge servers.
Recommended citation: Xu, Daliang, et al. "SoCFlow: Efficient and Scalable DNN Training on SoC-Clustered Edge Servers." ASPLOS (2024).
Published in ICSOC'2023', 2023
The first system that schedules DNN inference services on edge heterogeneous processors.
Recommended citation: Xu, Daliang, et al. "Niagara: Scheduling DNN Inference Services onHeterogeneous Edge Processors." ICSOC (2023).
Published in MobiCom'2022, 2022
The first system that enables highly resource-efficient on-device training by orchestrating mixed-precision training with on-chip Digital Signal Processor (DSP) offloading.
Recommended citation: Xu, Daliang, et al. "Mandheling: Mixed-Precision On-Device DNN Training with DSP Offloading." MobiCom (2022).