quantumlib/qsim

What it solves

qsim is a high-performance state-vector simulator for quantum circuits. It allows researchers and developers to simulate the evolution of quantum states over time by representing the state as a vector of complex-valued amplitudes and applying matrix-vector multiplication to simulate transformations.

How it works

As a full state-vector (Schrödinger) simulator, qsim computes all $2^n$ amplitudes of the state vector, where $n$ is the number of qubits. To achieve high performance, it employs several optimization techniques:

• Gate Fusion: Reduces the number of operations by combining multiple quantum gates into a single operation.
• Hardware Acceleration: Utilizes AVX/FMA instructions for vectorization, OpenMP for multithreading, and GPU support for faster computation.
• Precision: Uses single-precision arithmetic to speed up simulation.

Who it’s for

It is designed for users who need to simulate quantum circuits, including those using the C++ library directly or Python users via the qsimcirq interface, particularly those working with the Cirq framework for modeling Noisy Intermediate-Scale Quantum (NISQ) circuits.

Highlights

• High Performance: Highly tuned for vector arithmetic and multithreading on CPUs and GPUs.
• Cirq Integration: Includes a dedicated interface (qsimcirq) for simulating circuits written in Cirq.
• Proven Track Record: Used to produce landmark cross-entropy benchmark results published in Nature (2019).
• Flexible API: Available as both a C++ library and a Python package.