Getting Started

Choose Operation Module

diffqc provides multiple operation modules (aka. dense and sparse). These modules provide same functionality with different internal implementation.

If you don’t have special needs, we recommend to use dense module and to rename it for possible future replacement.

from diffqc import dense as op

Create Quantum State and Apply Quantum Gate

Initial |00...0> state can be created by op.zeros(nqubits, dtype).

import jax.numpy as jnp
from diffqc import dense as op

nqubits = 5
q = op.zeros(nqubits, jnp.complex64)

Quantum gate operations (e.g. PauliX, Hadamard, CNOT) take such quantum state, wire positions, and parameters (if exists), then return the new quantum state.

q = op.PauliX(q, (0,))
# |10000>

q = op.Hadamard(q, (1,))
# (|10000> + |11000>)/sqrt(2)

q = op.CNOT(q, (1, 2))
# (|10000> + |11100>)/sqrt(2)

Note

These quantum gate operations are executed immediately. Gate decomposition and/or optimization is out of scope in this project.

Expectation of Measurement

Expectation value at a wire position can be taken with corresponding function;

x = op.expectX(q, (0,)) # <q|X|q> at wire 0
y = op.expectY(q, (1,)) # <q|Y|q> at wire 1
z = op.expectZ(q, (2,)) # <q|Z|q> at wire 2

Convert to State-Vector Representation

State-Vector representation is a 1d vector with size of 2**(qubits). For 2 qubits, the values are probability amplitudes of [|00>, |01>, |10>, |11>].

sv = op.to_state(q)