qs8 is a quantum statevector simulator built in Python for small-scale projects. It was mostly an exercise in my own abilities in how the domains of linear algebra and quantum computing overlap, and only uses the numpy module as its core dependency.
In the future, I might adapt this project to JS for browser-based quantum simulation projects.
Quantum circuits are organized as a list of gates, where each entry is a timestep column. Index 0 is t0, index 1 is t1, etc. You can interpret the circuit by using matrix multiplication to evolve the statevector column by column, or build the entire circuit into a single unitary to then execute in one step. The single-stepping feature of this makes it easier to debug and figure out problems with state.
Pre-made gates include H, X, Y, Z, CX, CXl, and SWP. CX andCXl are the standard and lower version of the CNOT gate. Custom gates can be created as numpy arrays as seen in the current implementation of qs8.py:
# this has the first qb listed as control
CX = np.asarray([[1, 0, 0, 0],
[0, 0, 0, 1],
[0, 0, 1, 0],
[0, 1, 0, 0]])The core stipulation of this project is that there is currently no sugar on the multi-qubit operations. This means that qubits must be directly next to each other for a multi-qubit operation to take place.
For an example, let's create a circuit that initializes a GHZ state between 4 qubits.
from qs8 import *
qb = 4
# creating a circuit w/ 4 qubits
qc = QCirc(qb)After creating the quantum circuit with the associated amount of wires, or qubits, let's set the first couple of gates to test with.
# setting the gate on q0 in column 0 to X
qc.set_gate(X, [0], 0)
# setting the gate on q0 in column 1 to H
qc.set_gate(H, [0], 1)
The X gate runs a NOT on qubit 0, or q0, in column 0, and the H runs a Hadamard on q0 in column 1. Notice that the affected qubits are input as a list.
# create CX gates linking from q0 to qb-1 (q3)
for i in range(qb - 1):
qc.set_gate(CX, [i, i + 1], i + 2)
Now we use a for loop to add CX gates between q(i) and q(i+1)
# builds and runs circuit
qc.run_circuit(build=True)
# getting a dictionary of counts
counts = qc.get_counts(10000)
# print the circuit and counts
print(str(qc))
print(counts)
After the circuit is created, you can build the unitary while running by setting build=True. Additionally, you can interpret the circuit column by column by running qc.interpret_circuit() instead. You can get a dictionary of n shots by using the get_counts function on this circuit. Finally, the circuit has a __str__ method that prints the circuit in a readable format.
The output after running this script can be seen below:
q0 X-H-CX----
|
q1 ----o-CX--
|
q2 ------o-CX
|
q3 --------o-
{'15': 4986, '0': 5014}