Developer Ivan K. released ChartGPU, a new charting library that uses WebGPU to overcome CPU-bound limitations in Canvas2D and WebGL charting. The core innovation is moving all computation to the GPU, including LTTB downsampling as a compute shader and GPU-accelerated hit-testing for tooltips. This allows rendering 1 million points at 60fps with smooth zoom and pan, a significant leap for data-intensive applications.

Most charting libraries choke past 100K points because they rely on the CPU for data processing and rendering. By leveraging WebGPU's parallel compute power and instanced drawing, ChartGPU offloads these tasks entirely. The result is a library that supports line, area, bar, scatter, pie, and candlestick charts, all while maintaining interactive performance previously impossible in the browser.

Currently in its early stages (version 0.1.3), ChartGPU is MIT licensed and available on npm. It requires browsers with WebGPU support, like Chrome 113+ and Safari 18+, but not Firefox. The project includes a live demo showcasing the million-point capability, signaling a practical tool for developers working with large-scale real-time data visualization and analytics dashboards.