2 citations · 3 across the 14 of their papers we have counts for
15 papers
DB-SpMSpV: Dual-View Blocked Sparse Matrix-Sparse Vector Multiplication for Dynamic GPU Workloads
Xing Cong, Chenhao Xie, Rui Wang +3
Sparse Matrix-Sparse Vector Multiplication (SpMSpV) is a core primitive in graph traversal, sparse linear algebra, and sparse model inference. Its input vector is often dynamically…
Can LLMs Really Recover Microservice Failures? A Recovery-Aware Evaluation of Diagnosis-to-Action Reasoning
Jiaxing Qi, Zhongzhi Luan, Hongyu Zhang +5
Large language models (LLMs) are increasingly used to interpret operational evidence and assist incident response in cloud-native microservice systems. However, recovery-oriented u…
RATrain: A Resource-Aware Training Runtime for Large Language Models on Bandwidth-Constrained Heterogeneous Supercomputing Platforms
Yao Lu, Shiqing Ma, Zhongzhi Luan +5
Production heterogeneous supercomputing platforms are increasingly used to host large language model (LLM) training workloads. However, existing GPU-oriented training runtimes typi…
Bandwidth-Aware LLM Inference on Heterogeneous Many-Core Supercomputers
Yao Lu, Zhongzhi Luan, Gen Li +6
Large language model (LLM) inference is limited by high computational cost and memory bandwidth demands, making deployment on heterogeneous many-core processors challenging. Taking…
CB-SpMV:A Data Aggregating and Balance Algorithm for Cache-Friendly Block-Based SpMV on GPUs
Xing Cong, Fukai Sun, Yifan Chen +3
Sparse matrix-vector multiplication (SpMV) is crucial in computational science, engineering, and machine learning. Despite substantial efforts to improve SpMV performance on GPUs t…
AccelSync: Verifying Synchronization Coverage in Accelerator Pipeline Programs
Hangcheng An, Rui Wang, Depei Qian
AI accelerator operators are compiled into multi-stage pipeline programs where DMA, vector, matrix, and scalar units execute concurrently on shared on-chip buffers. A missing or mi…
RLHFSpec: Breaking the Efficiency Bottleneck in RLHF Training via Adaptive Drafting
Siqi Wang, Hailong Yang, Junjie Zhu +3
Reinforcement Learning from Human Feedback (RLHF) is an important fine-tuning technique for large language models (LLMs) and comprises three stages: generation, inference, and trai…
xGR: Efficient Generative Recommendation Serving at Scale
Qingxiao Sun, Tongxuan Liu, Shen Zhang +13
Recommendation system delivers substantial economic benefits by providing personalized predictions. Generative recommendation (GR) integrates LLMs to enhance the understanding of l…
PRAGMA: A Profiling-Reasoned Multi-Agent Framework for Automatic Kernel Optimization
Kelun Lei, Hailong Yang, Huaitao Zhang +5
Designing high-performance kernels requires expert-level tuning and a deep understanding of hardware characteristics. Recent advances in large language models (LLMs) have enabled a…
LOw-cOst yet High-Performant Sparse Matrix-Matrix Multiplication on Arm SME Architectures
Kelun Lei, Hailong Yang, Kaige Zhang +8
Sparse matrix-dense matrix multiplication (SpMM) is a critical kernel in both scientific computing and emerging graph learning workloads. The recent Armv9 architecture introduces S…
Beyond Window-Based Detection: A Graph-Centric Framework for Discrete Log Anomaly Detection
Jiaxing Qi, Chang Zeng, Zhongzhi Luan +5
Detecting anomalies in discrete event logs is critical for ensuring system reliability, security, and efficiency. Traditional window-based methods for log anomaly detection often s…
Quantum Machine Learning in Log-based Anomaly Detection: Challenges and Opportunities
Jiaxing Qi, Chang Zeng, Zhongzhi Luan +6
Log-based anomaly detection (LogAD) is the main component of Artificial Intelligence for IT Operations (AIOps), which can detect anomalous that occur during the system on-the-fly.…
FDLoRA: Personalized Federated Learning of Large Language Model via Dual LoRA Tuning
Yao Lu, Jiaxing QI, Zhongzhi Luan +4
Large language models (LLMs) have emerged as important components across various fields, yet their training requires substantial computation resources and abundant labeled data. It…
INSPIRIT: Optimizing Heterogeneous Task Scheduling through Adaptive Priority in Task-based Runtime Systems
Yiqing Wang, Xiaoyan Liu, Hailong Yang +5
As modern HPC computing platforms become increasingly heterogeneous, it is challenging for programmers to fully leverage the computation power of massive parallelism offered by suc…
Minions: Accelerating Large Language Model Inference with Aggregated Speculative Execution
Siqi Wang, Hailong Yang, Xuezhu Wang +10
Large language models (LLM) have recently attracted surging interest due to their outstanding capabilities across various domains. However, enabling efficient LLM inference is chal…