Memory Map Visualization

Memory Map Visualization

Back to Home

The Memory Map provides real-time visualization of the ByteBiota soup memory space, showing all organism instances, their species classifications, and current activities.

Features

  • Real-time Visualization: Live updates via WebSocket connection
  • Species Identification: Color-coded organisms by taxonomic classification
  • Activity Monitoring: Visual indicators for organism states
  • Interactive Exploration: Zoom, pan, and inspect individual organisms
  • Performance Optimized: Zero impact on simulator performance

Usage

Access the Memory Map via the web interface navigation or directly at /memory-map.

Controls

  • Zoom: Use mouse wheel or zoom buttons to change scale
  • Pan: Click and drag to move around the memory space
  • Select: Click on organisms to view detailed information
  • Filter: Use species and activity filters to focus on specific organisms
  • Export: Save memory snapshots as PNG images

Visual Elements

  • Organism Blocks: Colored rectangles representing organism memory ranges
  • Activity Indicators:
  • 🟢 Replicating
  • 🔴 Dying (high error count)
  • 🟣 Mutating
  • 🔵 Idle/Computing
  • Free Space: Gray areas representing unallocated memory
  • Heatmap: Optional overlay showing memory access patterns

Technical Details

API Endpoints

  • GET /api/memory/map: Get current memory map state (virtual memory layout in distributed mode)
  • WS /ws/memory-map: WebSocket for real-time updates (10 Hz update rate)

Note: In distributed mode, the memory map shows a virtual/logical memory space reconstructed from organism metadata across all workers. Addresses are virtual and do not correspond to actual worker memory locations.

Data Format

{
  "timestamp": 1234567890.123,
  "memory_size": 1048576,
  "free_ranges": [
    {"start": 0, "length": 1024}
  ],
  "organisms": [
    {
      "id": 1,
      "start_addr": 1024,
      "size": 512,
      "species": "Replicatus vulgaris",
      "activity": "executing",
      "current_ip": 1536,
      "energy": 45.2,
      "age": 1000,
      "error_count": 0
    }
  ]
}

Distributed Architecture {#distributed-arch}

In distributed ByteBiota, the memory map shows a virtual memory space reconstructed from organism metadata across all workers:

  1. No Worker Memory Access: Server uses existing StateAggregator organism data
  2. Virtual Addresses: Organisms assigned sequential addresses for visualization
  3. Real Metadata: Species, energy, age, activity states are actual values
  4. No Performance Impact: No additional worker communication required

This approach provides useful visualization without expensive memory synchronization.

Data Already Available {#data-available}

Workers send comprehensive organism updates including:
- Taxonomy classification (kingdom, phylum, genus, species)
- Energy levels and age
- Genome size and error counts
- All data needed for rich visualization

No additional worker requests needed - everything comes from existing execution result updates.

Data Export {#data-export}

The memory map system uses a dedicated MemoryMapService class to safely extract memory state data without impacting simulator performance. The service creates snapshots containing:

  • Organism locations and states
  • Memory allocation patterns
  • Species classifications
  • Activity indicators
  • Free memory ranges

Performance Considerations

  • Updates are throttled to prevent UI blocking
  • Large memory spaces show limited organism count for performance
  • Canvas rendering optimized for smooth 60fps interactions
  • WebSocket connection includes automatic reconnection