Data Virtualization Plugin

The Virtualization plugin is essential for maintaining 60 FPS performance when working with extremely large datasets. It uses intelligent Level-of-Detail (LOD) switching and dynamic downsampling to ensure the GPU only renders what is necessary for the current view.

Features

• ✅ Dynamic LOD: Automatically switches between full resolution and downsampled views based on zoom level.
• ✅ Optimized Strategies: Supports LTTB (Largest-Triangle-Three-Buckets) and Min-Max downsampling.
• ✅ Automatic Point Counting: Targets a specific number of points per pixel to maintain visual integrity while reducing load.
• ✅ Lazy-Load Integration: Works seamlessly with the Lazy Loading Plugin for a complete big-data pipeline.
• ✅ Memory Efficiency: Transparently manages data buffers to minimize overhead.

Basic Usage

import { createChart, PluginVirtualization } from 'scichart-engine';

const chart = createChart({
  container: document.getElementById('chart')!
});

// Initialize with LTTB strategy
await chart.use(PluginVirtualization({
  mode: 'lod',
  strategy: 'lttb',
  targetPoints: 'auto',
  pointsPerPixel: 2.5
}));

// The plugin will now automatically manage all series added to the chart
chart.addSeries({
  id: 'massive-series',
  data: { x: largeX, y: largeY } // e.g., 10 million points
});

API Reference

chart.virtualization

// Get performance stats for a series
const stats = chart.virtualization.getStats('massive-series');
console.log(`Original: ${stats.originalPoints}, Rendered: ${stats.renderedPoints}`);

// Invalidate cache and force re-downsample
chart.virtualization.invalidate('massive-series');

// Update configuration at runtime
chart.virtualization.updateConfig({
  strategy: 'minmax' // Switch to faster min-max for even larger sets
});

Configuration Options

Property	Type	Default	Description
mode	'lod' | 'bins' | 'hybrid'	'lod'	Virtualization mode.
strategy	'lttb' | 'minmax'	'lttb'	Downsampling algorithm.
targetPoints	number | 'auto'	'auto'	Max points to send to GPU.
pointsPerPixel	number	2	Detail density when targetPoints is 'auto'.
lodLevels	number[]	[1, 4, 8, 16]	Pre-calculated LOD factors.
includeSeries	string[]	all	Specific series to virtualize.

Virtualization Strategies

1. LTTB (Largest-Triangle-Three-Buckets)

Best for: General time-series where visual shape preservation is critical. LTTB is more computationally expensive but provides the best representation of peaks and valleys in the data.

2. Min-Max

Best for: Signals with very high frequency or noise where you want to see the "envelope" of the signal. Min-Max is extremely fast and ensures that outliers (highest and lowest points) are never missed.

Performance Tips

1. Use targetPoints: 'auto': This allows the plugin to adapt to the container size, preventing over-rendering on small screens.
2. Combine with Caching: Use the Caching Plugin to preserve downsampled buffers across view changes.
3. Wait for Idle: For background initialization, use the chart.virtualization.invalidate() method sparingly to avoid stuttering during interaction.

See Also

• Lazy Loading Plugin
• Caching Plugin
• Performance Guide