IBM Quantum Releases Open-Source “ffsim” Library for Specialized Fermionic Circuit Simulation

IBM Quantum researchers have released ffsim, an open-source Python library designed for the high-performance classical simulation of fermionic quantum circuits. Detailed in a recent arXiv preprint, the library addresses the computational challenges of modeling fundamental particles in atoms, molecules, and condensed matter by leveraging physical symmetries, such as total particle number and spin conservation. By restricting calculations to a subspace of the total Hilbert space, ffsim allows researchers to execute complex 64-qubit simulations on standard classical workstations, tasks that would otherwise require prohibitive memory resources on general-purpose simulators.

The structural design of ffsim utilizes a functional programming pattern based on NumPy, providing highly optimized backends for number-preserving operations including orbital rotations, Coulomb evolutions, and Trotter-Suzuki formulas. Furthermore, the library offers native integration with the Qiskit and PySCF ecosystems. Benchmark results indicate significant performance improvements: for a double-factorized Trotter simulation, ffsim demonstrated an 11x speedup over the Fermionic Quantum Emulator (FQE) and showed substantial gains in other fermionic Hamiltonian operations, effectively managing memory footprints that cause general-purpose engines like Qiskit Aer to plateau.

Source: quantumcomputingreport.com
Publication date: 12.06.2026
Author: Mohamed Abdel-Kareem