Memory-aware scheduling

Install

pip install matplotlib numpy
python main.py

Usage

from platforms.app import App

app = App()
app.read_input('your/sample/dag.json')

schedule, makespan, usage = app.schedule('heft_reserve', {'depth': 1})
print(makespan, usage)

Experiments

# install dependencies
pip install networkx wfcommons

# influence of different k
python experiments/depth.py
# minimum memory on random dags
python experiments/min_memory_statistics.py
# minimum memory on real world apps
python experiments/min_memory_real.py
# minimal makespan
python experiments/tradeoff.py

Samples

You can find different DAGs under samples directory.

• processor: same DAG but different # processors
• width: DAGs with different widths
• size: DAGS with same max width but different # nodes
• dependency: same DAG but different # edges

Real world applications

You can find real world applications in the paper under realworld_dax directory.

• Montage
• CyberShake
• Epigenomics
• SIGHT
• LIGO (Inspiral)