Running Variational Quantum Eigensolver with Qiskit Aer on AMD Instinct

2026-05-29 · Source: AMD ROCm Blogs · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Software Development & Engineering · Depth: Advanced, extended

Summary

GPU-accelerated Variational Quantum Eigensolver (VQE) simulations using Qiskit Aer on AMD Instinct MI300X GPUs with ROCm are demonstrated for quantum chemistry problems. The analysis benchmarks VQE on the LiH molecule across three basis sets: STO-3G (12 qubits), 6-31G (22 qubits), and 6-311G (32 qubits), comparing GPU versus CPU performance. For the 12-qubit STO-3G system, CPU and GPU performance is comparable (CPU slightly faster at 0.29s vs GPU 0.33s per evaluation). However, for the 22-qubit 6-31G system, the GPU achieves a 3.0x speedup (11.6s vs 35.1s per evaluation), reducing a full run to ~21.1 hours from ~2.7 days on CPU. At 32 qubits (6-311G), CPU simulation becomes impractical, while the GPU can still make progress, albeit at ~1.5 hours per optimization step. The VQE consistently achieves chemical accuracy (error below 1.59 × 10⁻³ Hartree) across all basis sets using the UCCSD ansatz and SLSQP optimizer.

Key takeaway

For research scientists evaluating quantum chemistry simulation platforms, AMD Instinct GPUs with Qiskit Aer and ROCm provide substantial acceleration for Variational Quantum Eigensolver (VQE) workloads. You should consider this stack to efficiently explore larger molecular systems, as it dramatically reduces simulation times for problems like LiH with 6-31G and 6-311G basis sets, enabling practical studies that are infeasible on CPU-only systems. This allows for faster algorithm development and validation.

Key insights

GPU-accelerated quantum circuit simulation significantly extends the practical scale for Variational Quantum Eigensolver (VQE) in quantum chemistry.

Principles

• VQE combines parameterized quantum circuits with classical optimizers.
• GPU acceleration improves VQE runtime for larger qubit counts.
• UCCSD ansatz and SLSQP optimizer achieve chemical accuracy for molecular problems.

Method

The VQE algorithm iteratively minimizes energy expectation values using a parameterized quantum circuit (ansatz) and a classical optimizer, converging to the ground state energy.

In practice

• Run VQE with Qiskit Aer on AMD Instinct GPUs via ROCm.
• Use UCCSD ansatz and SLSQP optimizer for accurate molecular ground state calculations.
• Track VQE experiments using MLflow for comprehensive analysis.

Topics

• Variational Quantum Eigensolver
• Qiskit Aer
• AMD Instinct GPUs
• ROCm
• Quantum Chemistry
• UCCSD Ansatz

Code references

• ROCm/rocm-blogs
• Qiskit/qiskit-aer
• qiskit-community/qiskit-nature

Best for: AI Scientist, Research Scientist, AI Hardware Engineer

Related on AIssential

• cuNNQS-SCI: A Fully GPU-Accelerated Framework for High-Performance Configuration Interaction Selection withNeural Network QQantum States — Fully GPU-accelerating NNQS-SCI overcomes CPU bottlenecks, enabling larger quantum system simulations with significant speedups.
• Quantum Feature Selection with Higher-Order Binary Optimization on Trapped-Ion Hardware — A quantum HUBO framework on trapped-ion hardware enhances feature selection by capturing higher-order data dependencies.
• Qilimanjaro Pushes Analog Quantum as AI Compute Demands Surge — Analog quantum computing offers a path to reduced errors and accelerated AI by avoiding extensive gate operations.
• How AI Will Change Quantum Computing | NVIDIA AI Podcast Ep. 294 — AI models, like NVIDIA Ising, are critical for accelerating quantum computing by addressing key challenges such as error correction and calibration.
• Alice & Bob expands Series B with investment from NVIDIA’s venture arm — The future of quantum computing is hybrid, integrating quantum and classical systems for real-world problem-solving.
• AMD Instinct™ GPUs MLPerf Inference v6.0 Submission — AMD's MI355X GPU and ROCm stack deliver competitive MLPerf Inference v6.0 results, especially for LLMs and multi-node setups.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by AMD ROCm Blogs.