|
| 1 | +import math |
| 2 | + |
| 3 | +import numpy as np |
| 4 | + |
| 5 | +from opensquirrel.common import Can1 |
| 6 | +from opensquirrel.gates import MultiQubitMatrixSemantic, Semantic, SingleQubitAxisAngleSemantic |
| 7 | + |
| 8 | + |
| 9 | +def extract_bits(x: int, bit_indices: [int]) -> int: |
| 10 | + """ |
| 11 | + Extract the bits of input at given indices, placing the bits in order from least to most significant. |
| 12 | + Equivalent to pext instruction. |
| 13 | +
|
| 14 | + Args: |
| 15 | + x: A non-negative integer from which one wants to extract the bits. |
| 16 | + bit_indices: The indices of the bits to extract, 0 being the least significant bit. |
| 17 | +
|
| 18 | + Returns: |
| 19 | + The extracted bits of x in order, as a non-negative integer. |
| 20 | +
|
| 21 | + Examples: |
| 22 | + >>> extract_bits(1, [0]) # 0b01 |
| 23 | + 1 |
| 24 | + >>> extract_bits(1111, [2]) # 0b01 |
| 25 | + 1 |
| 26 | + >>> extract_bits(1111, [5]) # 0b0 |
| 27 | + 0 |
| 28 | + >>> extract_bits(1111, [2, 5]) # 0b01 |
| 29 | + 1 |
| 30 | + >>> extract_bits(101, [1, 0]) # 0b10 |
| 31 | + 2 |
| 32 | + >>> extract_bits(101, [0, 1]) # 0b01 |
| 33 | + 1 |
| 34 | + """ |
| 35 | + result = 0 |
| 36 | + for i, bit_index in enumerate(bit_indices): |
| 37 | + result |= ((x & (1 << bit_index)) >> bit_index) << i |
| 38 | + |
| 39 | + return result |
| 40 | + |
| 41 | + |
| 42 | +def deposit_bits(x: int, bit_indices: [int]) -> int: |
| 43 | + """ |
| 44 | + Creates an integer whose bit values at given indices are taken from the bits of x, or 0 if they are not specified. |
| 45 | + Equivalently, takes the bits from x and places them at given indices in the result. |
| 46 | + Equivalent to pdep instruction. |
| 47 | +
|
| 48 | + Args: |
| 49 | + x: A non-negative integer giving the bit values. |
| 50 | + bit_indices: The indices where to deposit the bits of x in the result. |
| 51 | +
|
| 52 | + Returns: |
| 53 | + A bitstring whose bit values are taken from the bits of x. |
| 54 | +
|
| 55 | + Examples: |
| 56 | + >>> deposit_bits(0b0, [5]) # 0b000000 |
| 57 | + 0 |
| 58 | + >>> deposit_bits(0b1, [5]) # 0b100000 |
| 59 | + 32 |
| 60 | + >>> deposit_bits(0b000, [1, 2, 3]) # 0b0000 |
| 61 | + 0 |
| 62 | + >>> deposit_bits(0b001, [1, 2, 3]) # 0b0010 |
| 63 | + 2 |
| 64 | + >>> deposit_bits(0b011, [1, 2, 3]) # 0b0110 |
| 65 | + 6 |
| 66 | + >>> deposit_bits(0b0101, [1, 2, 3]) # 0b1010 |
| 67 | + 10 |
| 68 | + """ |
| 69 | + result = 0 |
| 70 | + for i, bit_index in enumerate(bit_indices): |
| 71 | + result |= ((x & (1 << i)) >> i) << bit_index |
| 72 | + |
| 73 | + return result |
| 74 | + |
| 75 | + |
| 76 | +def clear_bits(x: int, bit_indices: [int]) -> int: |
| 77 | + """ |
| 78 | + Clears given bits of input. |
| 79 | +
|
| 80 | + Args: |
| 81 | + x: A non-negative integer. |
| 82 | + bit_indices: Some bit indices to clear in x. |
| 83 | +
|
| 84 | + Returns: |
| 85 | + x with given bits reset to 0. |
| 86 | +
|
| 87 | + Examples: |
| 88 | + >>> clear_bits(0b1111, [1, 3]) # 0b0101 |
| 89 | + 5 |
| 90 | + """ |
| 91 | + result = x |
| 92 | + for index in bit_indices: |
| 93 | + result &= ~(1 << index) |
| 94 | + |
| 95 | + return result |
| 96 | + |
| 97 | + |
| 98 | +def get_expanded_matrix(semantic: Semantic, qubit_operands: [int], total_qubits: int) -> np.ndarray: |
| 99 | + """ |
| 100 | + Get the unitary matrix corresponding to the gate applied to those qubit operands. |
| 101 | + This can be used for, e.g., |
| 102 | + - testing, |
| 103 | + - permuting the operands of a multi-qubit gates, |
| 104 | + - simulating a circuit (simulation in this way is inefficient for large numbers of qubits). |
| 105 | +
|
| 106 | + Args: |
| 107 | + semantic: The semantic of the gate. |
| 108 | + qubit_operands: The qubit indices on which the gate operates. |
| 109 | + total_qubits: The total number of qubits. |
| 110 | +
|
| 111 | + Example: |
| 112 | + >>> X = SingleQubitAxisAngleSemantic((1, 0, 0), math.pi, math.pi / 2) |
| 113 | + >>> get_expanded_matrix(X, [1], 2).astype(int) # X q[1] |
| 114 | + array([[0, 0, 1, 0], |
| 115 | + [0, 0, 0, 1], |
| 116 | + [1, 0, 0, 0], |
| 117 | + [0, 1, 0, 0]]) |
| 118 | +
|
| 119 | + >>> CNOT = MultiQubitMatrixSemantic(np.array([[1, 0, 0, 0], [0, 1, 0, 0], [0, 0, 0, 1], [0, 0, 1, 0]])) |
| 120 | + >>> get_expanded_matrix(CNOT, [0, 2], 3) # CNOT q[0], q[2] |
| 121 | + array([[1, 0, 0, 0, 0, 0, 0, 0], |
| 122 | + [0, 0, 0, 0, 0, 1, 0, 0], |
| 123 | + [0, 0, 1, 0, 0, 0, 0, 0], |
| 124 | + [0, 0, 0, 0, 0, 0, 0, 1], |
| 125 | + [0, 0, 0, 0, 1, 0, 0, 0], |
| 126 | + [0, 1, 0, 0, 0, 0, 0, 0], |
| 127 | + [0, 0, 0, 0, 0, 0, 1, 0], |
| 128 | + [0, 0, 0, 1, 0, 0, 0, 0]]) |
| 129 | + >>> get_expanded_matrix(CNOT, [1, 2], 3) # CNOT q[1], q[2] |
| 130 | + array([[1, 0, 0, 0, 0, 0, 0, 0], |
| 131 | + [0, 1, 0, 0, 0, 0, 0, 0], |
| 132 | + [0, 0, 0, 0, 0, 0, 1, 0], |
| 133 | + [0, 0, 0, 0, 0, 0, 0, 1], |
| 134 | + [0, 0, 0, 0, 1, 0, 0, 0], |
| 135 | + [0, 0, 0, 0, 0, 1, 0, 0], |
| 136 | + [0, 0, 1, 0, 0, 0, 0, 0], |
| 137 | + [0, 0, 0, 1, 0, 0, 0, 0]]) |
| 138 | + """ |
| 139 | + if isinstance(semantic, SingleQubitAxisAngleSemantic): |
| 140 | + assert len(qubit_operands) == 1 |
| 141 | + |
| 142 | + which_qubit = qubit_operands[0] |
| 143 | + |
| 144 | + axis, angle, phase = semantic.axis, semantic.angle, semantic.phase |
| 145 | + result = np.kron( |
| 146 | + np.kron(np.eye(1 << (total_qubits - which_qubit - 1)), Can1(axis, angle, phase)), np.eye(1 << which_qubit) |
| 147 | + ) |
| 148 | + assert result.shape == (1 << total_qubits, 1 << total_qubits) |
| 149 | + return result |
| 150 | + |
| 151 | + assert isinstance(semantic, MultiQubitMatrixSemantic) |
| 152 | + |
| 153 | + # The convention is to write gate matrices with operands reversed. |
| 154 | + # For instance, the first operand of CNOT is the control qubit, and this is written as |
| 155 | + # 1, 0, 0, 0 |
| 156 | + # 0, 1, 0, 0 |
| 157 | + # 0, 0, 0, 1 |
| 158 | + # 0, 0, 1, 0 |
| 159 | + # which corresponds to control being q[1] and target being q[0], |
| 160 | + # since qubit #i corresponds to the i-th LEAST significant bit. |
| 161 | + qubit_operands.reverse() |
| 162 | + |
| 163 | + m = semantic.matrix |
| 164 | + |
| 165 | + assert m.shape == (1 << len(qubit_operands), 1 << len(qubit_operands)) |
| 166 | + |
| 167 | + result = np.zeros((1 << total_qubits, 1 << total_qubits), dtype=m.dtype) |
| 168 | + |
| 169 | + for input in range(1 << total_qubits): |
| 170 | + small_matrix_col_index = extract_bits(input, qubit_operands) |
| 171 | + |
| 172 | + col = m[:, small_matrix_col_index] |
| 173 | + |
| 174 | + for output, value in enumerate(col): |
| 175 | + large_output = clear_bits(input, qubit_operands) |
| 176 | + |
| 177 | + large_output |= deposit_bits(output, qubit_operands) |
| 178 | + |
| 179 | + result[large_output][input] = value |
| 180 | + |
| 181 | + assert result.shape == (1 << total_qubits, 1 << total_qubits) |
| 182 | + return result |
0 commit comments