| Task | Status |
|---|---|
| Construct or obtain datasets of N parallel time series | ✅ Done |
| Implement baseline approach to compute full correlation matrix | ✅ Done |
Measure runtime and peak memory usage via measure_correlation(data) |
✅ Done |
| Collect and store measured memory and runtime for different N | ✅ Done |
| Validate numerical correctness of computed correlations | ✅ Done |
Plot scaling results (plot_scaling() and plot_memory_breakdown()) |
✅ Done |
| Rigorous analysis of correlation matrix computation scaling with N | ✅ Done |
| Comparison of multiple memory methods against standard in-memory computation | ✅ Done |
| Brief report | ✅ Done |
| Main report (LaTeX PDF) | ✅ Done |
| Task | Status |
|---|---|
| Implement in psutil (RSS-based measurement) | ✅ Done |
| Try running the code without GarbageCollector | ✅ Done |
| Benchmark block-wise/tiled computation | ✅ Done |
| Streaming/chunked processing to avoid full N×N allocation | ✅ Done |
| On-the-fly standardization (z-score rows as needed) | ✅ Done |
| Approximate methods by sampling pairs of series | ✅ Done |
| Interactive UI dashboard (Streamlit) | ✅ Done |