WPF Real-Time Chart With 8 Billion Data Points

A WPF real-time chart built with LightningChart .NET can hold 8 billion data points in a live monitoring view while new data keeps arriving. This demo shows that running in both sweeping and scrolling modes, the two patterns used in instrumentation and monitoring software. The same engine drives WinForms, where performance is effectively identical.

What the demo shows

Two monitoring modes side by side. In sweeping mode the trace advances across a fixed view and overwrites the oldest data when it reaches the edge, the way a hospital monitor or an oscilloscope behaves. In scrolling mode the whole view moves so the newest data stays at the right edge.

Instead of a separate Y axis scale for every channel, the demo uses a shared MiniScale, also called a Gnomon. With many channels stacked in one view, individual axes consume most of the screen. A single reference scale gives back that space without losing the ability to read values.

Benchmark results

Measured in 2019 on LightningChart .NET 8.5. Current releases have not been re-benchmarked at this scale, so treat 8 billion as a floor rather than a ceiling. At these volumes memory is the binding constraint, not render time, so plan RAM around the data rather than around the chart.

How the chart is drawn

The trace is a rounded-corner polyline 1.5 pixels wide. Limit colouring changes the line to red or blue wherever values leave the nominal Y range, so an operator can see an excursion without reading the numbers. Rendering happens on the GPU rather than through the WPF visual tree, which is what keeps the frame rate stable while points are still arriving.

For a second look at the same engine under a different workload, see the scrolling line plot demo, which covers the one billion point case with a downloadable Visual Studio project.

Where real-time charts at this scale are used

• Medical monitoring. ECG, EEG and EMG, where several channels stream continuously and nothing can be dropped.
• Vibration and structural monitoring. High sample rate sensors on machinery, bridges and buildings.
• Motorsport and vehicle telemetry. Data loggers producing hundreds of channels per lap.
• Data acquisition and oscilloscopes. Instrument output where the waveform is the measurement.
• Audio waveform processing. Long recordings viewed end to end without resampling away detail.
• Trading. Thousands of simultaneous real-time feeds in one view.

Get the demo

The Visual Studio C# project for this demo is available on request. Contact us at [email protected], and we will send it over.

Common questions

How many data points can a WPF chart handle in real time?

Most WPF charting controls slow to an unusable frame rate between 100,000 and a few hundred thousand points, because they render through the CPU and the WPF visual tree. LightningChart .NET held 8 billion points in a live monitoring view in the 2019 benchmark above, with memory rather than render time as the limit.

What is the difference between sweeping and scrolling monitoring?

Sweeping keeps the view fixed and overwrites the oldest data when the trace reaches the right edge, like a hospital monitor. Scrolling moves the whole view so the newest data stays at the right edge and older data leaves the screen. Sweeping suits fixed-window monitoring, scrolling suits reviewing a moving history.

Does this work in WinForms as well as WPF?

Yes. The demo was built for WPF, but WinForms uses the same rendering engine and performs effectively the same.

Which .NET charting library is best for real-time data?

The deciding factor is where rendering happens. A library that creates a UI element per data point hits a wall early regardless of how the data is fed to it. GPU rendering scales with the graphics card instead of the UI thread, which is what makes billions of points in a live view possible. Our comparison of the best WPF charting libraries covers the options in more detail.

Want to see this on your own data? The full LightningChart .NET library is free for 30 days, with no card required.