Junyi Chen, Shihao Bai, Zaijun Wang, Siyu Wu, Chuheng Du, Hailong Yang, Ruihao Gong, Shengzhong Liu, Fan Wu, Guihai Chen. Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). 2025.

The exploration and application of Large Language Models (LLMs) is thriving. To reduce deployment costs, continuous batching has become an essential feature in current service frameworks. The effectiveness of continuous batching relies on an accurate estimate of the memory requirements of requests. However, due to the diversity in request output lengths, existing frameworks tend to adopt aggressive or conservative schedulers, which often result in significant overestimation or underestimation of memory consumption. Consequently, they suffer from harmful request evictions or prolonged queuing times, failing to achieve satisfactory throughput under strict Service Level Agreement (SLA) guarantees (a.k.a. goodput), across various LLM application scenarios with differing input-output length distributions. To address this issue, we propose a novel Past-Future scheduler that precisely estimates the peak memory resources required by the running batch via considering the historical distribution of request output lengths and calculating memory occupancy at each future time point. It adapts to applications with all types of input-output length distributions, balancing the trade-off between request queuing and harmful evictions, thereby consistently achieving better goodput. Furthermore, to validate the effectiveness of the proposed scheduler, we developed a high-performance LLM serving framework, LightLLM, that implements the Past-Future scheduler. Compared to existing aggressive or conservative schedulers, LightLLM demonstrates superior goodput, achieving up to 2-3× higher goodput than other schedulers under heavy loads. LightLLM is open source to boost the research in such direction (https://github.com/ModelTC/lightllm).

Computer simulation platforms offer an alternative solution by emulating complex systems in a controlled manner. However, existing Edge Computing (EC) simulators, as well as general-purpose vehicular network simulators, are not tailored for VEC and lack dedicated support for modeling the distinct access pattern, entity mobility trajectory and other unique characteristics of VEC networks. To fill this gap, this paper proposes VEC-Sim, a versatile simulation platform for in-depth evaluation and analysis of various service caching and computation offloading policies in VEC networks. VEC-Sim incorporates realistic mechanisms to replicate real-world access patterns, including service feature vector, vehicle mobility modeling, evolving service popularity, new service upload and user preference shifts, etc. Moreover, its modular architecture and extensive Application Programming Interfaces (APIs) allow seamless integration of customized scheduling policies and user-defined metrics. A comprehensive evaluation of VEC-Sim’s capabilities is undertaken in comparison to real-world ground truths. Results prove it to be accurate in reproducing classical scheduling algorithms and extremely effective in conducting case studies. • Vehicular edge network modeling establish foundation for simulator design • Modular architecture and rich APIs enable flexible simulator customization • Realistic-enhanced mechanisms replicate heterogeneous access patterns • In-depth experiments validate VEC-Sim’s feasibility and effectiveness

The increasing diversity of deep neural network (DNN) models and hardware platforms necessitates effective model profiling for high-performance inference deployment. Current DNN profiling tools suffer from either limited optimization insights due to the missing correlation between high-level DNN layer design and low-level hardware performance metrics, or prohibitive profiling overhead due to the large amount of performance measurement through hardware performance counters. Meanwhile, the roofline model has been widely used in the high-performance computing (HPC) domain for identifying performance bottlenecks and guiding optimizations. However, it lacks hierarchical (e.g., kernel/operator/layer), fine-grained, multi-platform support for profiling DNN models.

Social perception recommendation systems can effectively alleviate the user cold start problem by leveraging the side information of social networks. However, most social perception recommendation systems treat user relations as independently existing entities for learning, thereby overlooking potential connections between relations. Additionally, as the number of relations increases, it inevitably imposes a significant computational burden on the servers. To address these issues, we propose the Social perception recommendation based on relational clustering(SPRRC). SPRRC conducts relational clustering of social networks and projects knowledge graphs, effectively mining information about the similarity between relations. First, we cluster relationships in item knowledge graphs and social networks through unsupervised learning. After that, we use local weighted smoothing to aggregate the map information of the clustered items and social network virtual subgraphs respectively, and use the attention networks to learn the representation of users and items in the interactive bipartite graph. Finally, a large number of experiments have verified the high accuracy of our method, compared with the latest methods.