13 papers
CacheWise: Understanding Workloads and Optimizing KVCache Management for Efficiently Serving LLM Coding Agents
Shubham Tiwari, Tapan Chugh, Nash Rickert +3
Coding agents are a fast-growing LLM application, executing as long-running closed-loop sessions in which LLM generations alternate with external tool calls. Yet, unlike chat workl…
C2CServe: Leveraging NVLink-C2C for Elastic Serverless LLM Serving on MIG
Shutian Luo, Ali Zafar Sadiq, Rui Yang +4
Modern LLM serving is increasingly serverless in shape: large model catalogs, long-tail invocations, and multi-tenant demand. Existing GPU serving systems face a tradeoff: dedicate…
Unifying Sparse Attention with Hierarchical Memory for Scalable Long-Context LLM Serving
Zihan Zhao, Baotong Lu, Shengjie Lin +8
Long-context LLM serving is bottlenecked by the cost of attending over ever-growing KV caches. Dynamic sparse attention promises relief by accessing only a small, query-dependent s…
DeepCompile: A Compiler-Driven Approach to Optimizing Distributed Deep Learning Training
Masahiro Tanaka, Du Li, Umesh Chand +3
The rapid growth of deep learning models has increased the demand for efficient distributed training strategies. Fully sharded approaches like ZeRO-3 and FSDP partition model param…
MorphServe: Efficient and Workload-Aware LLM Serving via Runtime Quantized Layer Swapping and KV Cache Resizing
Zhaoyuan Su, Zeyu Zhang, Tingfeng Lan +4
Efficiently serving large language models (LLMs) under dynamic and bursty workloads remains a key challenge for real-world deployment. Existing serving frameworks and static model…
Straggler Tolerant and Resilient DL Training on Homogeneous GPUs
Zeyu Zhang, Haiying Shen
Despite the popularity of homogeneous GPU-based deep learning (DL) training, the prevalence, causes and impact of stragglers and the effectiveness of existing straggler mitigation…