The Engine Behind
RadarSimX
A high-performance C++/CUDA radar simulation engine combining OpenMP multi-threading and GPU parallelization. Used directly by RadarSimPy, and via RadarSimLib by RadarSimM and RadarSimApp.
Built for Performance and Precision
A dual-acceleration architecture delivering accuracy and speed for complex radar scenarios.
CPU Parallelization
- Multi-threaded execution via OpenMP
- Efficient simulation on standard hardware
- Supported on Windows, Linux, and macOS
- No special hardware required
GPU Acceleration
- Massive parallelization via CUDA 13
- Fully exploits NVIDIA GPU processing power
- Supported on Windows and Linux
- Requires compatible NVIDIA driver
Comprehensive Radar Modeling
- CW, FMCW, PMCW, pulse, and arbitrary waveforms
- Advanced modulation — CDM, FDM, DDM, TDM
- Phase noise modeling
- Multi-channel Tx/Rx with antenna patterns
High-Fidelity EM Simulation
- Ray-tracing for 3D target & environment modeling
- Point target & RCS simulation
- Interference simulation
- LiDAR point cloud generation
Platform Support
Available on all supported platforms. Scales automatically with available CPU cores for fast simulation on any machine.
Available on Windows and Linux with NVIDIA GPUs. Delivers orders-of-magnitude speedup for large-scale scenarios.
| Platform | CPU (OpenMP) | GPU (CUDA 13) |
|---|---|---|
| Windows | ✓ | ✓ |
| Linux (Ubuntu 22.04 / 24.04) | ✓ | ✓ |
| macOS (Intel) | ✓ | — |
| macOS (Apple Silicon) | ✓ | — |
Commercial License
RadarSimCpp is distributed under a commercial license. Pre-compiled libraries are bundled with RadarSimPy, RadarSimM, and RadarSimApp. Source code access requires a separate licensing agreement.
Organizations requiring full source code access for customization or deep integration can contact us for a detailed quote. For most use cases, the pre-built interfaces (RadarSimPy, RadarSimM, RadarSimApp) provide complete access to the engine’s capabilities without needing the source.