Time Slot: Track 1 10:20-10:50 Language: English Speaker: Mr. Erik Welch | NVIDIA | USA
Have you ever wondered how those data scientists at Facebook and LinkedIn make friend recommendations? Or how epidemiologists track down patient zero in an outbreak? Or how to make graph algorithms really, really fast? If so, then this talk is for you.
NetworkX is the most popular library in Python for graph theory and applied networkx science thanks to its extensive API and beginner-friendly documentation. You should use it. However, what should you do when your graph data becomes too large or NetworkX becomes too slow? Simple: use an accelerated NetworkX backend!
NetworkX 3 can now be used with other highly-tuned graph libraries:
- cuGraph, which is part of NVIDIA RAPIDS to accelerate data science on GPUs
- python-graphblas, which uses linear algebra to write elegent, efficient graph algorithms for the CPU and GPU
We will show workflows that achieve 100 to 10000+ times speedup using these backends!
Mr. Erik Welch
I work on open-source at NVIDIA: cuGraph, NetworkX, python-graphblas, Dask, RAPIDS, toolz, etc.