Summary of Gnnbench: Fair and Productive Benchmarking For Single-gpu Gnn System, by Yidong Gong and Pradeep Kumar
GNNBENCH: Fair and Productive Benchmarking for Single-GPU GNN System
by Yidong Gong, Pradeep Kumar
First submitted to arxiv on: 5 Apr 2024
Categories
- Main: Machine Learning (cs.LG)
- Secondary: Distributed, Parallel, and Cluster Computing (cs.DC)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This research paper proposes a standardized benchmarking platform for Graph Neural Network (GNN) systems called GNNBench. The authors argue that the lack of a unified benchmark has led to fundamental design and evaluation pitfalls in the community, which are overlooked. To address this issue, they present a protocol for exchanging tensor data, support custom classes in system APIs, and enable automatic integration with various deep learning frameworks like PyTorch and TensorFlow. The authors demonstrate the importance of GNNBench by integrating several GNN systems, revealing measurement issues that require attention from the community. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary GNNBench is a new way to test and compare Graph Neural Networks. Right now, there’s no standard way to do this, which means that some important problems are being overlooked. The researchers created a platform called GNNBench that makes it easy to design and evaluate GNN systems. They show how using GNNBench can help identify issues with measuring the performance of these systems. |
Keywords
* Artificial intelligence * Attention * Deep learning * Gnn * Graph neural network
