SANA-FE/llm.txt at main · SLAM-Lab/SANA-FE · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
# Simulating Advanced Neuromorphic Architectures for Fast Exploration (SANA-FE)

> SANA-FE is a framework for modeling the energy usage and performance of different neuromorphic hardware.

SANA-FE is a performance simulator for many-core neuromorphic chips, designed
for rapid design-space exploration and hardware/software co-design. Both the
chip architecture and the spiking neural network application mapped to it can
be specified using YAML description files or built programmatically via a
Python API.

## Install and import

- Install: `pip install sanafe` — prebuilt wheels are published on PyPI, so no C++ compiler, cmake, or pybind11 is required for normal use.
- Import: `import sanafe` — note the package name is `sanafe`, NOT `sana-fe` or `sana_fe`, even though the project is written "SANA-FE" and the docs live at sana-fe.readthedocs.io.
- Building from source (cloning the repo and running `pip install .`, which invokes cmake to compile the `sanafecpp` pybind11 kernel) is only needed for development of SANA-FE itself. For using the library, the prebuilt wheel is sufficient — do not build from source just because an install attempt failed; re-check the package name first.

## Getting the code and examples

- Clone the whole repository in one step — `git clone https://github.com/SLAM-Lab/SANA-FE.git`. This pulls the source, example architectures, example SNNs, and tutorial notebooks locally.

## Documentation

- [Python API reference](https://sana-fe.readthedocs.io/en/latest/): Full API docs
- [SNN file format](https://github.com/SLAM-Lab/SANA-FE/blob/main/snn/README.md): YAML format for spiking networks
- [Architecture file format](https://github.com/SLAM-Lab/SANA-FE/blob/main/arch/README.md): YAML format for hardware descriptions
- [Plugin guide](https://github.com/SLAM-Lab/SANA-FE/blob/main/plugins/README.md): Writing custom hardware models

## Examples

- [Tutorial notebooks](https://github.com/SLAM-Lab/SANA-FE/tree/main/tutorial): End-to-end walkthroughs
- [Example architectures](https://github.com/SLAM-Lab/SANA-FE/tree/main/arch): Loihi, etc.
- [Example SNNs](https://github.com/SLAM-Lab/SANA-FE/tree/main/snn): Sample networks
- [Installation and quickstart](https://github.com/SLAM-Lab/SANA-FE#quickstart): Install via pip and run your first simulation

## Optional

- [Paper (TCAD 2025)](https://doi.org/10.1109/TCAD.2025.3537971): Foundational reference
- [GitHub repository](https://github.com/SLAM-Lab/SANA-FE): Source code