yangkatiezhao/awesome-gnn-systems

A list of awesome GNN systems.

Awesome Graph Neural Network Systems

A list of awesome systems for graph neural network (GNN). If you have any comment, please create an issue or pull request.

Contents

• Open Source Libraries
• Papers
  • Survey Papers
  • GNN Libraries
  • GNN Kernels
  • GNN Compilers
  • Distributed GNN Training Systems
  • Training Systems for Scaling Graphs
  • Quantized GNNs
  • GNN Dataloaders
  • GNN Training Accelerators
  • GNN Inference Accelerators
• Contribute

Open Source Libraries

• PyG: Graph Neural Network Library for PyTorch
• DGL: Python Package Built to Ease Deep Learning on Graph
• Graph Nets: Build Graph Nets in Tensorflow
• Euler: A Distributed Graph Deep Learning Framework
• StellarGraph: Machine Learning on Graphs
• Spektral: Graph Neural Networks with Keras and Tensorflow 2
• PGL: An Efficient and Flexible Graph Learning Framework Based on PaddlePaddle
• CogDL: An Extensive Toolkit for Deep Learning on Graphs
• DIG: A Turnkey Library for Diving into Graph Deep Learning Research
• Jraph: A Graph Neural Network Library in Jax
• Graph-Learn: An Industrial Graph Neural Network Framework
• DeepGNN: a Framework for Training Machine Learning Models on Large Scale Graph Data

Papers

Survey Papers

Venue	Title	Affiliation	Link	Source
arXiv 2022	Distributed Graph Neural Network Training: A Survey	BUPT	[paper]
arXiv 2022	Parallel and Distributed Graph Neural Networks: An In-Depth Concurrency Analysis	ETHZ	[paper]
CSUR 2022	Computing Graph Neural Networks: A Survey from Algorithms to Accelerators	UPC	[paper]

GNN Libraries

Venue	Title	Affiliation	Link	Source
JMLR 2021	DIG: A Turnkey Library for Diving into Graph Deep Learning Research	TAMU	[paper]	[code]
arXiv 2021	CogDL: A Toolkit for Deep Learning on Graphs	THU	[paper]	[code]
CIM 2021	Graph Neural Networks in TensorFlow and Keras with Spektral	Università della Svizzera italiana	[paper]	[code]
arXiv 2019	Deep Graph Library: A Graph-Centric, Highly-Performant Package for Graph Neural Networks	AWS	[paper]	[code]
VLDB 2019	AliGraph: A Comprehensive Graph Neural Network Platform	Alibaba	[paper]	[code]
arXiv 2019	Fast Graph Representation Learning with PyTorch Geometric	TU Dortmund University	[paper]	[code]
arXiv 2018	Relational Inductive Biases, Deep Learning, and Graph Networks	DeepMind	[paper]	[code]

GNN Kernels

Venue	Title	Affiliation	Link	Source
MLSys 2022	Understanding GNN Computational Graph: A Coordinated Computation, IO, and Memory Perspective	THU	[paper]	[code]
HPDC 2022	TLPGNN: A Lightweight Two-Level Parallelism Paradigm for Graph Neural Network Computation on GPU	GW	[paper]
IPDPS 2021	FusedMM: A Unified SDDMM-SpMM Kernel for Graph Embedding and Graph Neural Networks	Indiana University Bloomington	[paper]	[code]
SC 2020	GE-SpMM: General-purpose Sparse Matrix-Matrix Multiplication on GPUs for Graph Neural Networks	THU	[paper]	[code]
ICCAD 2020	fuseGNN: Accelerating Graph Convolutional Neural Network Training on GPGPU	UCSB	[paper]	[code]
IPDPS 2020	PCGCN: Partition-Centric Processing for Accelerating Graph Convolutional Network	PKU	[paper]

GNN Compilers

Venue	Title	Affiliation	Link	Source
MLSys 2022	Graphiler: Optimizing Graph Neural Networks with Message Passing Data Flow Graph	ShanghaiTech	[paper]	[code]
EuroSys 2021	Seastar: Vertex-Centric Programming for Graph Neural Networks	CUHK	[paper]
SC 2020	FeatGraph: A Flexible and Efficient Backend for Graph Neural Network Systems	Cornell	[paper]	[code]

Distributed GNN Training Systems

Venue	Title	Affiliation	Link	Source
VLDB 2022	Sancus: Staleness-Aware Communication-Avoiding Full-Graph Decentralized Training in Large-Scale Graph Neural Networks	HKUST	[paper]	[code]
MLSys 2022	BNS-GCN: Efficient Full-Graph Training of Graph Convolutional Networks with Partition-Parallelism and Random Boundary Node Sampling	Rice, UIUC	[paper]	[code]
MLSys 2022	Sequential Aggregation and Rematerialization: Distributed Full-batch Training of Graph Neural Networks on Large Graphs	Intel	[paper]	[code]
WWW 2022	PaSca: A Graph Neural Architecture Search System under the Scalable Paradigm	PKU	[paper]
ICLR 2022	PipeGCN: Efficient Full-Graph Training of Graph Convolutional Networks with Pipelined Feature Communication	Rice	[paper]	[code]
ICLR 2022	Learn Locally, Correct Globally: A Distributed Algorithm for Training Graph Neural Networks	PSU	[paper]	[code]
arXiv 2021	Distributed Hybrid CPU and GPU training for Graph Neural Networks on Billion-Scale Graphs	AWS	[paper]
SC 2021	DistGNN: Scalable Distributed Training for Large-Scale Graph Neural Networks	Intel	[paper]	[code]
SC 2021	Efficient Scaling of Dynamic Graph Neural Networks	IBM	[paper]
CLUSTER 2021	2PGraph: Accelerating GNN Training over Large Graphs on GPU Clusters	NUDT	[paper]
OSDI 2021	$P^3$: Distributed Deep Graph Learning at Scale	MSR	[paper]
OSDI 2021	Dorylus: Affordable, Scalable, and Accurate GNN Training with Distributed CPU Servers and Serverless Threads	UCLA	[paper]	[code]
arXiv 2021	GIST: Distributed Training for Large-Scale Graph Convolutional Networks	Rice	[paper]
EuroSys 2021	FlexGraph: A Flexible and Efficient Distributed Framework for GNN Training	Alibaba	[paper]
EuroSys 2021	DGCL: An Efficient Communication Library for Distributed GNN Training	CUHK	[paper]	[code]
SC 2020	Reducing Communication in Graph Neural Network Training	UC Berkeley	[paper]	[code]
VLDB 2020	G$^3$: When Graph Neural Networks Meet Parallel Graph Processing Systems on GPUs	NUS	[paper]	[code]
IA3 2020	DistDGL: Distributed Graph Neural Network Training for Billion-Scale Graphs	AWS	[paper]	[code]
MLSys 2020	Improving the Accuracy, Scalability, and Performance of Graph Neural Networks with Roc	Stanford	[paper]	[code]
arXiv 2020	AGL: A Scalable System for Industrial-purpose Graph Machine Learning	Ant Financial Services Group	[paper]
ATC 2019	NeuGraph: Parallel Deep Neural Network Computation on Large Graphs	PKU	[paper]

Training Systems for Scaling Graphs

Venue	Title	Affiliation	Link	Source
EuroSys 2023	Marius++: Large-Scale Training of Graph Neural Networks on a Single Machine	UW–Madison	[paper]	[code]
VLDB 2022	ByteGNN: Efficient Graph Neural Network Training at Large Scale	ByteDance	[paper]
ICML 2022	GraphFM: Improving Large-Scale GNN Training via Feature Momentum	TAMU	[paper]	[code]
ICML 2021	GNNAutoScale: Scalable and Expressive Graph Neural Networks via Historical Embeddings	TU Dortmund University	[paper]	[code]
OSDI 2021	GNNAdvisor: An Adaptive and Efficient Runtime System for GNN Acceleration on GPUs	UCSB	[paper]	[code]

Quantized GNNs

Venue	Title	Affiliation	Link	Source
Neurocomputing 2022	EPQuant: A Graph Neural Network Compression Approach Based on Product Quantization	ZJU	[paper]	[code]
ICLR 2022	EXACT: Scalable Graph Neural Networks Training via Extreme Activation Compression	Rice	[paper]	[code]
PPoPP 2022	QGTC: Accelerating Quantized Graph Neural Networks via GPU Tensor Core	UCSB	[paper]	[code]
CVPR 2021	Binary Graph Neural Networks	ICL	[paper]	[code]
CVPR 2021	Bi-GCN: Binary Graph Convolutional Network	Beihang University	[paper]	[code]
EuroMLSys 2021	Learned Low Precision Graph Neural Networks	Cambridge	[paper]
World Wide Web 2021	Binarized Graph Neural Network	UTS	[paper]
ICLR 2021	Degree-Quant: Quantization-Aware Training for Graph Neural Networks	Cambridge	[paper]	[code]
ICTAI 2020	SGQuant: Squeezing the Last Bit on Graph Neural Networks with Specialized Quantization	UCSB	[paper]	[code]

GNN Dataloaders

Venue	Title	Affiliation	Link	Source
NSDI 2023	BGL: GPU-Efficient GNN Training by Optimizing Graph Data I/O and Preprocessing	ByteDance	[paper]
MLSys 2022	Accelerating Training and Inference of Graph Neural Networks with Fast Sampling and Pipelining	MIT	[paper]	[code]
EuroSys 2022	GNNLab: A Factored System for Sample-based GNN Training over GPUs	SJTU	[paper]	[code]
KDD 2021	Global Neighbor Sampling for Mixed CPU-GPU Training on Giant Graphs	UCLA	[paper]
PPoPP 2021	Understanding and Bridging the Gaps in Current GNN Performance Optimizations	THU	[paper]	[code]
VLDB 2021	Large Graph Convolutional Network Training with GPU-Oriented Data Communication Architecture	UIUC	[paper]	[code]
TPDS 2021	Efficient Data Loader for Fast Sampling-Based GNN Training on Large Graphs	USTC	[paper]	[code]
SoCC 2020	PaGraph: Scaling GNN Training on Large Graphs via Computation-aware Caching	USTC	[paper]	[code]
arXiv 2019	TigerGraph: A Native MPP Graph Database	UCSD	[paper]

GNN Training Accelerators

Venue	Title	Affiliation	Link	Source
ISCA 2022	Graphite: Optimizing Graph Neural Networks on CPUs Through Cooperative Software-Hardware Techniques	UIUC	[paper]
ISCA 2022	Hyperscale FPGA-as-a-service architecture for large-scale distributed graph neural network	Alibaba	[paper]
arXiv 2021	GCNear: A Hybrid Architecture for Efficient GCN Training with Near-Memory Processing	PKU	[paper]
DATE 2021	ReGraphX: NoC-enabled 3D Heterogeneous ReRAM Architecture for Training Graph Neural Networks	WSU	[paper]
TCAD 2021	Rubik: A Hierarchical Architecture for Efficient Graph Learning	Chinese Academy of Sciences	[paper]
FPGA 2020	GraphACT: Accelerating GCN Training on CPU-FPGA Heterogeneous Platforms	USC	[paper]	[code]

GNN Inference Accelerators

Venue	Title	Affiliation	Link	Source
JAIHC 2022	DRGN: a dynamically reconfigurable accelerator for graph neural networks	XJTU	[paper]
DAC 2022	GNNIE: GNN Inference Engine with Load-balancing and Graph-specific Caching	UMN	[paper]
IPDPS 2022	Understanding the Design Space of Sparse/Dense Multiphase Dataflows for Mapping Graph Neural Networks on Spatial Accelerators	GaTech	[paper]	[code]
arXiv 2022	FlowGNN: A Dataflow Architecture for Universal Graph Neural Network Inference via Multi-Queue Streaming	GaTech	[paper]
CICC 2022	StreamGCN: Accelerating Graph Convolutional Networks with Streaming Processing	UCLA	[paper]
HPCA 2022	Accelerating Graph Convolutional Networks Using Crossbar-based Processing-In-Memory Architectures	HUST	[paper]
HPCA 2022	GCoD: Graph Convolutional Network Acceleration via Dedicated Algorithm and Accelerator Co-Design	Rice, PNNL	[paper]	[code]
arXiv 2022	GenGNN: A Generic FPGA Framework for Graph Neural Network Acceleration	GaTech	[paper]
DAC 2021	DyGNN: Algorithm and Architecture Support of vertex Dynamic Pruning for Graph Neural Networks	Hunan University	[paper]
DAC 2021	BlockGNN: Towards Efficient GNN Acceleration Using Block-Circulant Weight Matrices	PKU	[paper]
DAC 2021	TARe: Task-Adaptive in-situ ReRAM Computing for Graph Learning	Chinese Academy of Sciences	[paper]
ICCAD 2021	G-CoS: GNN-Accelerator Co-Search Towards Both Better Accuracy and Efficiency	Rice	[paper]
MICRO 2021	I-GCN: A Graph Convolutional Network Accelerator with Runtime Locality Enhancement through Islandization	PNNL	[paper]
arXiv 2021	ZIPPER: Exploiting Tile- and Operator-level Parallelism for General and Scalable Graph Neural Network Acceleration	SJTU	[paper]
TComp 2021	EnGN: A High-Throughput and Energy-Efficient Accelerator for Large Graph Neural Networks	Chinese Academy of Sciences	[paper]
HPCA 2021	GCNAX: A Flexible and Energy-efficient Accelerator for Graph Convolutional Neural Networks	GWU	[paper]
APA 2020	GNN-PIM: A Processing-in-Memory Architecture for Graph Neural Networks	PKU	[paper]
ASAP 2020	Hardware Acceleration of Large Scale GCN Inference	USC	[paper]
DAC 2020	Hardware Acceleration of Graph Neural Networks	UIUC	[paper]
MICRO 2020	AWB-GCN: A Graph Convolutional Network Accelerator with Runtime Workload Rebalancing	PNNL	[paper]
arXiv 2020	GRIP: A Graph Neural Network Accelerator Architecture	Stanford	[paper]
HPCA 2020	HyGCN: A GCN Accelerator with Hybrid Architecture	UCSB	[paper]

Contribute

We welcome contributions to this repository. To add new papers to this list, please update JSON files under ./res/papers/. Our bots will update the paper list in README.md automatically. The citations of newly added papers will be updated within one day.