- Add memory parameter to task launches based on used data
- Fix concurrent read/writes in 1D Stencil
- Use bandwidth estimate to configure and estimate data transfer costs in microseconds
- Build spawn_id to task_id dictionary
- Get location of data movement tasks based on runtime output
- Use runtime output to verify data movement correctness
- Compute total data movement cost
- Use thresholded weights
- Refactor viz into anaylze.py
- Pull apart plot function to build_graph and plot_graph
- This was anaylsis can be run separately