quDiet

quDiet

A high performance simulator that scales and eats qudits for lunch. Read More

Installation

A simple pip install from the github repository will install the package in your system.

Install with pip

$ pip install git+https://github.com/LegacYFTw/QuDiet

Install from source

$ git clone https://github.com/LegacYFTw/QuDiet && cd QuDiet
$ pip3 install .

Examples

Here are some simple examples.

We start by importing the package:

from qudiet.core.quantum_circuit import QuantumCircuit

The class QuantumCircuit is the main interface representing the abstraction over the complicated calculations.

A 3-Qudit Example

Let's start with this 3-qudit example, where is a qubit whereas and are qutrits.

First, we start with initializing a blank circuit with the class QuantumCircuit.

qc = QuantumCircuit(

    # Represents the dimension of each quantum register
    qregs=[2, 3, 3],


    # Represents the initial states
    init_states=[0, 0, 0],
)

At a glance, the class qudiet.core.quantum_circuit.QuantumCircuit looks quite similar to Qiskit's QuantumCircuit class, with an exception of the qregs defining the qubit dimension of the registers.

Now that we have a blank circuit, we can place quantum gates like this,

# Places the CX gate with control qudit q0, q1 and target qudit q2
qc.cx([0, 1], 2)

# The measure_all definition represents the end of the definition of a QuantumCircuit 
qc.measure_all()

Now that we are done with the definition of our QuantumCircui, we can execute the circuit with the .run() definition.

qc.run()

Backends

QuDiet comes with 4 different kinds of backend.

# The NumPy Backend is the default Backend
from qudiet.core.backend.NumpyBackend import NumpyBackend

# Sparse Backend works great when there are no H gates
from qudiet.core.backend.SparseBackend import SparseBackend

# Experimental - GPU

# CUDA Backend is the GPU equivalent of the NumPy Backend
from qudiet.core.backend.CUDABackend import CUDABackend

# CUDA Sparse Backend is the GPU equivalent of the Sparse Backend
from qudiet.core.backend.CUDASparseBackend import CUDASparseBackend

These Backends can be declared when initializing the circuit with the class qudiet.core.quantum_circuit.QuantumCircuit using the argumentbackend=NumpyBackend.

qc = QuantumCircuit(
    qregs=[2, 3, 3],
    init_states=[0, 0, 0],

    # Here goes the backend specification
    backend=NumpyBackend,
)

Run from QASM

A typical qasm looks like

# This is the File Header
# The File Header section is compulsory

# Defining a 3 Qudit circuit
.qubit 3
qudit x0 (2)
qudit x1 (3)
qudit x2 (3)

# The Gate composition will go in this
# section, between the .begin and the .end
.begin

# Your Gates Here ...
Toffoli x0, x1, x2

.end

To run a circuit from qasm, we need to import a few definitions,

from qudiet.qasm.qasm_parser import circuit_from_qasm, parse_qasm

The circuit_from_qasm definition creates the circuit from the qasm str. Whereas, the parse_qasm definition creates the circuit directly from the qasm file definition.

# Loading the circuit from the qasm string
circuit = circuit_from_qasm(
    '''
    # This is the File Header
    
    .qubit 3
    qudit x0 (2)
    qudit x1 (3)
    qudit x2 (3)

    .begin
    Toffoli x0, x1, x2
    .end
    '''
)

# Now we can run the circuit however we want.
circuit.run()

The parse_qasm definition can be used as such.

# Loading the circuit from a qasm file
filename = "test.qasm"
circuit = parse_qasm(filename, SparseBackend)

# Now we can run the circuit however we want.
result = circuit.run()

Adding Arbitary Gates

Custom arbitary gates can be defined as a child class of qudiet.circuit_library.ArbitaryGate.

from qudiet.circuit_library import ArbitaryGate

class GateXYZ(ArbitaryGate):
    def __init__(self, *args, **kwargs):
        super().__init__(*args, **kwargs)
        self._unitary = np.array(
            [
                [1, 0, 0], 
                [0, 1, 0], 
                [0, 0, 1], 
            ]
        )

Once the gate id defined, it can be applied to a circuit as such,

# Defining a Quantum Circuit
qc = QuantumCircuit(
    qregs=[2, 3, 4, 3, 2],
    init_states=[1, 2, 2, 0, 1],
)

# The definition QuantumCircuit.gate() applies an Arbitary gate 
# QuantumCircuit.gate( ArbitaryGate , Register/s )
qc.gate(GateXYZ, 3) # Here, we are applying the gate `GateXYZ` to the fourth register, i.e. qreg-3

Bibliography

@article{chatterjee2022qudiet,
  title={QuDiet: A Classical Simulation Platform for Qubit-Qudit Hybrid Quantum Systems},
  author={Chatterjee, Turbasu and Das, Arnav and Bala, Subhayu Kumar and Saha, Amit and Chattopadhyay, Anupam and Chakrabarti, Amlan},
  journal={arXiv preprint arXiv:2211.07918},
  year={2022}
}