The RAPIDS cuGraph library is a collection of GPU accelerated graph algorithms that process data found in GPU DataFrames.
Research: Infectious Disease Modeling, Monsoon 2020
Guided by: Prof. Kishore Kothapalli, Prof. Sriram Pemmaraju
See Colaboratory for notebook, but it takes about 15min to setup Rapid's
cuGraph onto it, before starting. BlazingSQL has cuGraph already setup
onto it, but you need to write your own notebook (it has no share feature).
With BlazingSQL you need to copy data
as well and main.py
into its
environment.
## OUTPUT
## Small graph: 5 vertices, 5 edges
Contact Graph: (5, 2) (5, 3)
Infectable Graph: (5, 2) (2, 3)
Path graph: (5, 2) (3, 3)
src dst wt
1 0 2 10
2 3 0 9
distance vertex predecessor
0 9.0 3 0
1 10.0 2 0
2 0.0 0 -1
## OUTPUT
## Large graph: 50000 vertices, 249272 edges
Contact Graph: (50000, 2) (249272, 3)
Infectable Graph: (50000, 2) (112381, 3)
Path graph: (50000, 2) (48619, 3)