heavyai/heavydb

What it solves

HeavyDB is designed to handle massive datasets that are too large or complex for traditional databases, specifically by leveraging the parallel processing power of GPUs to enable high-speed SQL queries and real-time visualization of geospatial and relational data.

How it works

It operates as an in-memory, column-store SQL relational database that runs directly on GPUs. The platform integrates several components:

• HeavyDB Core: The GPU-accelerated database engine.
• Immerse UI: A front-end for creating dashboards and advanced visualizations.
• GPU Renderer: A back-end system that generates visual elements like heatmaps, polygons, and rasters based on Vega, which are then composited over maps.
• HeavyConnect: A data ingestion layer that uses foreign tables to analyze external data (CSV, Parquet, S3, ODBC) without complex ETL processes.
• HeavyRF: A specialized module for simulating RF propagation for wireless network planning.

Who it’s for

It is built for developers and data analysts who have access to GPU resources and need to process complex geospatial data, perform high-speed simulations, or build real-time data dashboards.

Highlights

• GPU-Native Architecture: Built from the ground up to run on GPUs for maximum performance.
• Integrated Visualization: Includes the Immerse UI and a GPU-accelerated renderer for complex mapping and imagery.
• Full Platform Open Source: Now includes enterprise-grade features previously restricted to paid versions.
• Simplified Data Access: Supports direct analysis of external files and databases via HeavyConnect.