feagi-burst-engine
High-performance neural processing unit (NPU) for FEAGI burst cycle execution.
Overview
The burst engine processes neural activity in discrete time steps:
- Synaptic propagation
- Neural dynamics (membrane potential updates)
- Firing detection and management
- Optional GPU acceleration (WGPU, CUDA)
Installation
[]
= "2.0"
# With GPU support
= { = "2.0", = ["gpu"] }
# With CUDA support (NVIDIA only)
= { = "2.0", = ["cuda"] }
Usage
use RustNPU;
let mut npu = new?;
npu.load_connectome?;
npu.process_burst?;
Features
gpu- Cross-platform GPU acceleration via WGPUcuda- NVIDIA CUDA accelerationall-gpu- Enable all GPU backends
Performance
- 50-100x faster than Python implementations
- Supports 30Hz+ burst frequency
- Tested with millions of neurons
Part of the FEAGI ecosystem.