GPU simulation and real-time visualization.
Making data and ideas computable, connected and visible.
Correlate Visuals develops high-performance simulation and visualization tools. The work centres on the GPU: simulations that run entirely in compute shaders, renderers that show them as they happen, and data modelled so its structure can be seen.
Core technologies
- C++ and Vulkan for cross-platform engines
- GLSL for rendering and GPU compute
- CUDA and OpenCL for parallel computation
- Python for data processing and visualization
- TouchDesigner for interactive and live visuals
Specializations
- GPU simulation: cellular automata and agent-based systems in compute shaders
- Real-time graphics and Vulkan engine architecture
- Interactive visuals for installations and live performance
- Knowledge modelling and data visualization
- Emergent systems: from Conway's Game of Life to simulated societies whose cells produce, trade and build.
- Scientific data made visible: astronomical images, stellar diagrams and point clouds.
- Software that lasts: deterministic results, bounded memory and programs that run for years.
- Knowledge as graphs: structuring texts and ideas so their connections can be explored.
Vulpen is a Vulkan engine for real-time visuals and GPU compute, where programs are graphs of nodes written in C++ and GLSL. The engine places every Vulkan barrier from the graph's connections, swaps in code changes while the project runs, and can replay any session from its command log.
Linux and Windows · Mozilla Public License 2.0 · github.com/CorrelateVisuals/Vulpen-GPU
The other repositories here hold TouchDesigner components, CUDA and OpenCL kernel collections, and knowledge-modelling projects.


