Graph the planet: Wrangling GPU graph dataframes with GFQL
summary
GFQL is an open source, dataframe-native graph query language that runs in Python on CPU or NVIDIA GPUs and claims up to 42X speedups on real-world graphs and a Graph 500 third place finish on first submission. It loads tabular source/destination CSVs or QDF GPU dataframes, runs neighborhood and clustering queries, and integrates with PyData, Arrow and GPU libraries; examples include billion‑dollar lawsuits, cybersecurity incident response, clickstream analytics and mining the Bluesky firehose, all usable after a simple pip install.
AI-written summary · about
chapters (15)
- 0:00Graphistry and graphs intro
- 1:14Twitter COVID graphs
- 7:57Graphs vs spreadsheets
- 9:28Graph data use cases
- 15:52GFQL introduction
- 19:40Why GFQL doesn’t need a new database
- 21:45GFQL with dataframes and GPUs
- 25:02GPU vs CPU for graphs
- 29:44GFQL benchmarking
- 31:36Graphs in one line with GFQL
- 33:24LLMs with GFQL
- 36:35GFQL in Splunk AI challenge
- 40:13Encouraging graph thinking
- 44:09Start with GFQL today
- 46:14Q&A
related talks
41mGraphGen: Conducting Graph Analytics over Relational Databases
30mGraph Data Science
37mGraph Computing at Scale
46mGraph Databases: Talking about your Data Relationships with Python
29mTigerGraph, Connecting the World
18mRethinking Data: A Data Scientist’s First Dive into GQLs
29mAll about graphs in Database
49mModern Graph Databases