How faisalahmed3301/academic works
This is a single-script research experiment comparing classical and learned page-replacement policies on a synthetic trace with a deliberate distribution shift. The workflow runs from CLI parameters through workload generation, learned-policy training, policy simulation, and result reporting. The graph separates the experiment’s domain stages and its internal policy/feature layer. The source excerpt is partial, so broader training and reporting links rely on README documentation; no unsampled implementation details are inferred.
Main entry points: main.py Experiment runner · decision_tree_rules.txt Tree rules · results Result charts · summary_table.csv Summary table
Components 14
Workload design
- Shifted trace
main.py - Locality phase
main.py - Bursty phase
main.py
Learned policy
- Eviction features
main.py - OPT labels
main.py - Decision tree
main.py - Tree rules
results/decision_tree_rules.txt
Evaluation and outputs
- Experiment runner
main.py - Replacement policies
main.py - Phase metrics
main.py - Result charts
results - Summary table
results/summary_table.csv
Other components
- Researcher
- CLI entry
main.py
Connections 15
- Researcher → CLI entry: runs
README.md - CLI entry → Experiment runner: starts
main.py - Experiment runner → Shifted trace: generates
main.py - Shifted trace → Locality phase: includes phase
README.md - Shifted trace → Bursty phase: includes phase
README.md - Experiment runner → Eviction features: trains from
README.md - Eviction features → OPT labels: labeled by
main.py - Experiment runner → Decision tree: trains
README.md - Decision tree → Eviction features: uses
README.md - Experiment runner → Replacement policies: simulates
main.py - Replacement policies → Decision tree: predicts with
main.py - Experiment runner → Phase metrics: computes
README.md - Phase metrics → Summary table: writes
README.md - Experiment runner → Result charts: generates
README.md - Decision tree → Tree rules: exports
main.py