Huayou Su

TileGEMM: Boosting the Performance of GEMM on AMX-Powered CPUs by Exploiting Data Reuse

Kangkang Chen, Huayou Su, Menghan Jia, Yong Dou

General Matrix Multiplication (GEMM) is the cornerstone of high-performance computing and deep learning. Its efficiency significantly influences the performance of applications ranging from large language models to scientific simulations. Intel Advanced Matrix Extensions (AMX) significantly boost matrix operations throughput, yet existing implementations …

cn (code pays fourni par la source)

AMXStencil: Boosting the Performance of Stencil Computations on AMX-Powered CPUs via Fusion

Zitong An, Kangkang Chen, Huayou Su, Jinwei Xu et autres

cn (code pays fourni par la source)

DOA: Dataflow Optimization for Attention on Multi-core DSPs with Three-Level Memory Hierarchy

Shengwei Li, Zhiquan Lai, Shuai Xu, Shun Ouyang et autres

cn (code pays fourni par la source)

JSQKV: Joint Sparsification and Quantization for KV-Cache Compression and Decode Acceleration

Hao Zhang, Xiaoli Gong, Haoran Li, Huayou Su et autres

cn (code pays fourni par la source)

SparX: Cache-Aware Sparse Matrix Multiplication Kernels on AMX-Based CPUs

Ziqi Wang, Dongsheng Li, Huayou Su

cn (code pays fourni par la source)

Model Profiling-Based Model Routing Method for Mathematical Reasoning

Zhaoxie Xu, Xu Zhang, Xudong Gong, Dawei Feng et autres

cn (code pays fourni par la source)

Eliminating Redundant Feature Tests in Decision Tree and Random Forest Inference on SQL Predicates

Mingxi Liu, Zhengyuan Ding, Chenyang Zhang, Qingfeng Pan et autres

In-database prediction queries that apply machine learning (ML) pipelines to perform data analysis are prevalent in many applications. Since data stored in databases is typically tabular, tree-based models are particularly well-suited and thus widely adopted for such tasks. When ML inference with …

cn, us (code pays fourni par la source)

Quantitative Analysis and Performance Optimization of Graph Neural Networks on Multi-core CPUs

Huayou Su, Xi Yang, Zitong An, Yong Dou et autres

Graph Neural Networks (GNNs) are becoming increasingly popular in graph data processing due to their excellent performance in feature extraction on graph datasets. Compared to GPUs, CPUs are more widely accessible and serve as a practical platform for GNN inference. However, achieving …

cn (code pays fourni par la source)

Pro-Prophet: A Systematic Load Balancing Method for Efficient Parallel Training of Large-Scale MoE Models

Wei Wang, Zhiquan Lai, Dongsheng Li, Shengwei Li et autres

The size of deep learning models has been increasing to enhance model quality. The linear increase in training computation budgets with model size means that training an extremely large-scale model is exceedingly time-consuming. Recently, the Mixture of Experts (MoE) has drawn significant …

cn (code pays fourni par la source)

Heterogeneous Feature-Aware Graph Neural Network for Tabular Data

Hongxiao Fei, Jinqi Hu, Liu Yang, Tingxuan Chen et autres

cn (code pays fourni par la source)

ORE: An Offline Redundancy Elimination System for GNN Acceleration

Ziqi Wang, Yongquan Fu, Huayou Su

cn (code pays fourni par la source)

REAMP: A Redundancy Elimination System for AMP-GNN Acceleration

Ziqi Wang, Yongquan Fu, Huayou Su

cn (code pays fourni par la source)