# GeNN

GeNN (GPU-enhanced Neuronal Networks) is a code-generation-based simulation environment for spiking neural network models, compiling network descriptions to native NVIDIA CUDA or AMD HIP code for GPU execution.

GeNN simulates spiking neural networks by generating native GPU code from a network's description rather than interpreting it at runtime, targeting NVIDIA CUDA or AMD HIP so the same modeling approach can run on either vendor's hardware.

Building it requires a C++ compiler appropriate to the platform, Visual Studio's Desktop development with C++ workload on native Windows, or GCC 7.5 or later plus libffi's development package on Linux or Windows with WSL2. For NVIDIA GPUs, a compatible CUDA toolkit must be installed and selected through the CUDA_PATH environment variable, matched to a compatible C++ compiler version since the newest compiler is not always compatible with the newest CUDA release; for AMD GPUs, HIP is installed separately with HIP_PATH and HIP_PLATFORM set accordingly. GeNN is used in computational neuroscience research to simulate large spiking neural network models on GPU hardware.

Website: http://genn-team.github.io/