A few days ago I started experimenting with gpui for the second time. The first time I just tried out the examples and glanced at the code, this time around I’m trying to build an interactive and customizable network graph library. I was able to get a proof-of-concept up relatively quickly but it’s poorly made. Time to refine this into a library and see just how well gpui will perform with large graphs.
gpug progress
- refactor gpug into a library
- learn how cargo examples work & refactor the original gpug binary into examples
- explore ‘hornet’ for dynamic sparse graph analysis
- determine how to test performance for the different examples we will build
- test performance
- to be continued…
Reading & Research
Graphs stand or fall by their choice of nodes and edges.
- Started researching graph data structures on 10/16/2025.
- Started researchin graph visualization approaches on ?/?/?.
- Started research graph analysis approaches on ?/?/?.
Hornet: An Efficient Data Structure for Dynamic Sparse Graphs and Matrices on GPUs
10/16/2025
I’ll start reviewing the code for the hornet paper today, its ideas of fast updates+traversal sound valuable. Perhaps there will be something here I can use or adapt. This will likely take me a few days.
The NetworkX approach
research me and translate to rust + compare performance…