3 papers
cs.CV2026
AdaFocus: Adaptive Relevance-Diversity Sampling with Zero-Cache Look-back for Efficient Long Video Understanding
Xiao Yang, Yingzhe Ma, Haoxuan Yu +2
Long video understanding is heavily bottlenecked by a rigid one-shot paradigm: existing methods either densely encode videos at prohibitive memory and latency costs, or aggressivel…
cs.DC2025
Torpor: GPU-Enabled Serverless Computing for Low-Latency, Resource-Efficient Inference
Minchen Yu, Ao Wang, Dong Chen +8
Serverless computing offers a compelling cloud model for online inference services. However, existing serverless platforms lack efficient support for GPUs, hindering their ability…
cs.DC2025
SGDRC: Software-Defined Dynamic Resource Control for Concurrent DNN Inference on NVIDIA GPUs
Yongkang Zhang, Haoxuan Yu, Chenxia Han +7
Cloud service providers heavily colocate high-priority, latency-sensitive (LS), and low-priority, best-effort (BE) DNN inference services on the same GPU to improve resource utiliz…