Performance Testing Scripts - SM vs ST Comparison

This repository contains performance analysis scripts and data for comparing SharedMatrix (SM) vs SharedTree (ST) implementations in the Fluid Framework.

📊 Key Performance Charts

Primary Time Performance Analysis

• time_period_with_error_bar_chart - 🏆 Main time performance diagram
  • Measures Period (ns/op) - the core metric for operation speed
  • Includes error bars for statistical reliability
  • Uses logarithmic scale to handle 10x-1000x performance differences
  • Shows SM is 10-100x faster than ST across all operations

Primary Memory Performance Analysis

• memory_bar_chart - 🏆 Main memory performance diagram
  • Measures Heap Used Average (bytes) - actual memory consumption
  • Includes error bars for measurement reliability
  • Shows both allocation and deallocation patterns (positive/negative values)
  • Shows SM uses more predictable, smaller memory footprints

📈 Additional Analysis Scripts

Time Analysis

• time_elapsed_time_bar_chart - Total elapsed time comparison
• time_relative_margin_error_chart - Measurement uncertainty analysis

Memory Analysis

• memory_elapsed_time_chart - Memory operation timing
• memory_heap_stddev_chart - Memory usage variability
• memory_relative_margin_error_chart - Memory measurement uncertainty
• memory_split_charts - Separate small/large scale memory operations

📁 Data Files

• time.csv - Time performance data (Period, Elapsed Time, Margin of Error)
• memory.csv - Memory performance data (Heap Usage, Standard Deviation, Margin of Error)

🚀 Usage

Prerequisites

pip install pandas plotly kaleido

Running the Scripts

# Primary time analysis
python time_period_with_error_bar_chart

# Primary memory analysis
python memory_bar_chart

# Additional analyses
python [script_name]

Output Files

Each script generates:

• .png - High-resolution charts
• .html - Interactive visualizations
• .json - Raw chart data