A Benchmarking Framework for Superconductor-Based Neural Networks
Welcome to the official website repository for SuperSIM, a benchmarking framework designed to evaluate the performance of neural networks on superconducting Josephson devices, specifically focusing on AQFP-based PIM architectures.
🔗 Live Site: cryo-lab.github.io/SuperSIM-Website
SuperSIM provides in-depth architecture-level simulations and performance assessments to support the development of superconducting PIM (Processing-in-Memory) chips.
✅ Supports single-bit & multi-bit PIM architectures
✅ Provides energy, latency, and area estimations
✅ Integrates circuit-level models for precise benchmarking
✅ Offers PyTorch compatibility for neural network development
✅ Includes cycle-accurate simulations for performance evaluation
📄 Reference Paper: SuperSIM: A Comprehensive Benchmarking Framework for Neural Networks Using Superconductor Josephson Devices
To preview this Jekyll-based website on your local machine:
git clone https://github.com/cryo-lab/SuperSIM-Website.git
cd SuperSIM-Website