PSAL-POSTECH/PyTorchSim

PyTorchSim is a Comprehensive, Fast, and Accurate NPU Simulation Framework

53
/ 100
Established

This project helps evaluate the performance and efficiency of custom AI accelerator hardware designs, specifically Neural Processing Units (NPUs). It takes your existing PyTorch deep learning models (like ResNet, BERT, GPT-2) and produces detailed reports on how they would perform on a simulated NPU architecture. This is designed for hardware architects and system engineers developing new NPU designs.

Use this if you are designing or optimizing custom NPU hardware and need to accurately predict the performance and resource utilization of deep learning workloads before physical fabrication.

Not ideal if you are solely a deep learning practitioner focused on model development and training, or if you need to simulate general-purpose CPUs or GPUs.

NPU-design hardware-simulation AI-accelerator-development deep-learning-systems hardware-architecture
No Package No Dependents
Maintenance 10 / 25
Adoption 9 / 25
Maturity 16 / 25
Community 18 / 25

How are scores calculated?

Stars

97

Forks

17

Language

Python

License

MIT

Last pushed

Mar 13, 2026

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/ml-frameworks/PSAL-POSTECH/PyTorchSim"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.