Quantum computing simulation tools available now on Lakeshore!
Introduction
Quantum computing simulation tools available now on Lakeshore!
Quantum computing simulation tools available now on Lakeshore!
Quantum computing simulation is one of the most computationally demanding workloads in modern research. ACER has installed and validated three leading quantum simulation frameworks on Lakeshore’s GPU nodes:
- CUDA-Q 0.12.0 (NVIDIA)
- Qiskit Aer 0.15.1 (IBM)
- PennyLane 0.42.3 (Xanadu)
GPU-accelerated quantum simulation is available now to UIC researchers through a single pip install command on both the H100 NVL (batch_gpu2) and A100-SXM4 MIG (batch_gpu) partitions.
Each framework is designed for a different use case. Qiskit is the most widely used quantum SDK and the natural starting point for new users. PennyLane is built around automatic differentiation and is best suited for quantum machine learning and variational algorithms. CUDA-Q is an HPC-native platform optimized for large-scale GPU simulation with native multi-GPU support.
Benchmarking on Lakeshore shows GPU simulation is 22,000x faster than CPU at 21 qubits, with 31-qubit circuits completing in under one second on the H100. Above 25 qubits, GPU is not just faster — it is the only viable option.
Learn more, including step-by-step installation instructions, Slurm job templates, and benchmark results, on our Confluence knowledge base pages https://confluence.acer.uic.edu/display/KB/High+Performance+Computing.
If you have specific needs related to quantum workflows or scaling guidance, contact ACER at acer@uic.edu.
*Please make sure you access Confluence links on the UIC Network or through the UIC VPN.