6 papers · 1 filter
CommBench: Can LLMs Write Correct and Efficient GPU Communication Code?
Shuang Ma, Yuyi Li, Yihan Zhang +12
Training and serving large language models (LLMs) rely heavily on high-performance GPU communication, yet implementing efficient GPU communication primitives requires deep expertis…
UCCL-Zip: Lossless Compression Supercharged GPU Communication
Shuang Ma, Chon Lam Lao, Zhiying Xu +8
The rapid growth of large language models (LLMs) has made GPU communication a critical bottleneck. While prior work reduces communication volume via quantization or lossy compressi…
UCCL-EP: Portable Expert-Parallel Communication
Ziming Mao, Yihan Zhang, Chihan Cui +9
Mixture-of-Experts (MoE) workloads rely on expert parallelism (EP) to achieve high GPU efficiency. State-of-the-art EP communication systems such as DeepEP demonstrate strong perfo…
ConServe: Fine-Grained GPU Harvesting for LLM Online and Offline Co-Serving
Yifan Qiao, Shu Anzai, Shan Yu +10
Large language model (LLM) serving demands low latency and high throughput, but high load variability makes it challenging to achieve high GPU utilization. In this paper, we identi…
Locality-aware Fair Scheduling in LLM Serving
Shiyi Cao, Yichuan Wang, Ziming Mao +10
Large language model (LLM) inference workload dominates a wide variety of modern AI applications, ranging from multi-turn conversation to document analysis. Balancing fairness and…
NEO: Saving GPU Memory Crisis with CPU Offloading for Online LLM Inference
Xuanlin Jiang, Yang Zhou, Shiyi Cao +2
Online LLM inference powers many exciting applications such as intelligent chatbots and autonomous agents. Modern LLM inference engines widely rely on request batching to improve i…