This repo contains examples that will be included in Legate's documentation (https://github.com/nv-legate)
Run madelbrot.py
using different backends and see how easy it is to get performance improvements
To run using NumPy
python mandelbrot.py
Use 8 OpenMP threads to compute the Mandelbrot set
LEGATE_TEST=1 legate --omps 1 --ompthreads 8 --regmem 10000 --sysmem 10000 ./mandelbrot.py
Use 1 GPU to compute the Mandelbrot set
LEGATE_TEST=1 legate --gpus 1 --fbmem 10000 ./mandelbrot.py
Use 4 CPUs threads to compute the Mandelbrot set.
LEGATE_TEST=1 legate --cpus 4 --sysmem 10000 ./mandelbrot.py
(Link to FAQs on how to time, what these input arguments mean etc.)