The parity table PT describes the phase factor in the phase polynomial intermediate representation. Technically, the parity table on its own is not an intermediate representation as it further requires the phase angles α and parity matrixP to describe (CNOT, RZ) circuits (phase polynomials),
∣x⟩=e−i2α(1−2⋅PTx)∣Px⟩.
Sometimes, the phase factor 2α(1−2⋅PTx) itself is called the phase polynomial, which adds confusion to the terminology.
We refer to phase polynomial intermediate representation for a full overview.
We go through each gate of the circuit and note down its action on the binary input state ∣x1,x2,x3,x3⟩. A CNOT gate adds the control value onto the target value.
A RZ gate collects the current parity, which is the logical value of that qubit. E.g. the first CNOT alters the first qubit value to x1⊕x2. The corresponding parity that we collect at the first RZ gate is (1,1,0,0), as we have n=4 qubits in total, and its angle θ1. We continue doing so throughout the circuit and collect the parities θ1(x1⊕x2), θ2(x1⊕x2⊕x3), and θ3(x1⊕x3⊕x4), which we can summarize in the corresponding parity table
PT=⎝⎛111110011001⎠⎞,
and angle vector θ=(θ1,θ2,θ3).
The overall accumulated phase of the circuit is given by
[1] "Phase polynomials synthesis algorithms for NISQ architectures and beyond", Vivien Vandaele, Simon Martiel, Timothée Goubault de Brugière arXiv:2104.00934, 2021
Cite this page
@misc{PennyLane-ParityTable,
title = "Parity Table",
author = "Korbinian Kottmann",
year = "2025",
howpublished = "\url{https://pennylane.ai/compilation/parity-table}"
}