microsoft/qdk
Publicmirrored fromhttps://github.com/microsoft/qdkAvailable
samples/notebooks/parallel_teleport.ipynb
210lines · modecode
| 1 | { |
| 2 | "cells": [ |
| 3 | { |
| 4 | "cell_type": "markdown", |
| 5 | "id": "f94615ed", |
| 6 | "metadata": {}, |
| 7 | "source": [ |
| 8 | "# Experimenting with Parallel Teleport on Neutral Atom Device\n", |
| 9 | "\n", |
| 10 | "This notebook shows how you can use the `NeutralAtomDevice` class to visualize and simulate a scaled, parallel teleport sample program.\n", |
| 11 | "\n", |
| 12 | "First, set up the environment." |
| 13 | ] |
| 14 | }, |
| 15 | { |
| 16 | "cell_type": "code", |
| 17 | "execution_count": null, |
| 18 | "id": "ae94a027", |
| 19 | "metadata": {}, |
| 20 | "outputs": [], |
| 21 | "source": [ |
| 22 | "from qdk import init, TargetProfile, code, Result\n", |
| 23 | "from qdk.qsharp import compile, circuit\n", |
| 24 | "from qdk.simulation import NeutralAtomDevice, NoiseConfig\n", |
| 25 | "from qdk.widgets import Circuit, Histogram\n", |
| 26 | "\n", |
| 27 | "init(target_profile=TargetProfile.Base)" |
| 28 | ] |
| 29 | }, |
| 30 | { |
| 31 | "cell_type": "markdown", |
| 32 | "id": "df5d6387", |
| 33 | "metadata": {}, |
| 34 | "source": [ |
| 35 | "Next, we use Q# to define a `ParalellTeleport` that uses distinct sets of qubits to perform a measurement deferred teleportation. Using different qubits rather than the same qubits should allow later device-specific compilation steps to identify parallelism and run those teleports across the qubits at the same time." |
| 36 | ] |
| 37 | }, |
| 38 | { |
| 39 | "cell_type": "code", |
| 40 | "execution_count": null, |
| 41 | "id": "0d435592", |
| 42 | "metadata": { |
| 43 | "vscode": { |
| 44 | "languageId": "qsharp" |
| 45 | } |
| 46 | }, |
| 47 | "outputs": [], |
| 48 | "source": [ |
| 49 | "%%qsharp\n", |
| 50 | "\n", |
| 51 | "operation ParallelTeleport(instances : Int) : Result[] {\n", |
| 52 | " // Partitions the teleport instances across rows and columns on the machine\n", |
| 53 | " let cols = if instances >= 12 { 36 } else { instances * 3 };\n", |
| 54 | " let rows = (instances + 11) / 12; // 1 to 12 = 1, 13 to 24 = 2, etc.\n", |
| 55 | "\n", |
| 56 | " use qubits = Qubit[instances * 3];\n", |
| 57 | " mutable results : Result[] = [];\n", |
| 58 | "\n", |
| 59 | " for i in 0..instances-1 {\n", |
| 60 | " let rowId = i / 12;\n", |
| 61 | " let colId = (i % 12) * 3;\n", |
| 62 | " let idx = colId + (rowId * 36);\n", |
| 63 | "\n", |
| 64 | " // Prep state on msg qubit\n", |
| 65 | " if i % 4 == 1 {\n", |
| 66 | " X(qubits[idx]);\n", |
| 67 | " } elif i % 4 == 2 {\n", |
| 68 | " H(qubits[idx]);\n", |
| 69 | " } elif i % 4 == 3 {\n", |
| 70 | " SX(qubits[idx]);\n", |
| 71 | " }\n", |
| 72 | "\n", |
| 73 | " Teleport(qubits[idx], qubits[idx + 1], qubits[idx + 2]);\n", |
| 74 | "\n", |
| 75 | " // Reverse state prep on Bob's qubit\n", |
| 76 | " if i % 4 == 1 {\n", |
| 77 | " X(qubits[idx + 2]);\n", |
| 78 | " } elif i % 4 == 2 {\n", |
| 79 | " H(qubits[idx + 2]);\n", |
| 80 | " } elif i % 4 == 3 {\n", |
| 81 | " X(qubits[idx + 2]);\n", |
| 82 | " SX(qubits[idx + 2]);\n", |
| 83 | " }\n", |
| 84 | " results += [MResetZ(qubits[idx + 2])];\n", |
| 85 | " ResetAll([qubits[idx], qubits[idx + 1]]);\n", |
| 86 | " }\n", |
| 87 | "\n", |
| 88 | " return results;\n", |
| 89 | "}\n", |
| 90 | "\n", |
| 91 | "operation Teleport(msg : Qubit, alice : Qubit, bob : Qubit) : Unit {\n", |
| 92 | " // Create some entanglement that we can use to send our message.\n", |
| 93 | " H(alice);\n", |
| 94 | " CNOT(alice, bob);\n", |
| 95 | "\n", |
| 96 | " // Encode the message into the entangled pair.\n", |
| 97 | " CNOT(msg, alice);\n", |
| 98 | " H(msg);\n", |
| 99 | "\n", |
| 100 | " CNOT(alice, bob);\n", |
| 101 | " Controlled Z([msg], bob);\n", |
| 102 | "}" |
| 103 | ] |
| 104 | }, |
| 105 | { |
| 106 | "cell_type": "markdown", |
| 107 | "id": "d86e19f4", |
| 108 | "metadata": {}, |
| 109 | "source": [ |
| 110 | "You can verify the circuit for two parallel teleports using the `Circuit` widget." |
| 111 | ] |
| 112 | }, |
| 113 | { |
| 114 | "cell_type": "code", |
| 115 | "execution_count": null, |
| 116 | "id": "b815acd4", |
| 117 | "metadata": {}, |
| 118 | "outputs": [], |
| 119 | "source": [ |
| 120 | "Circuit(circuit(code.ParallelTeleport, 2))" |
| 121 | ] |
| 122 | }, |
| 123 | { |
| 124 | "cell_type": "markdown", |
| 125 | "id": "c342d7e9", |
| 126 | "metadata": {}, |
| 127 | "source": [ |
| 128 | "Now, let's visualize 120 instances of teleportation in parallel on the `NeutralAtomDevice`. This will decompose the program to the device's native gate set and schedule the operations with movement, showing how it might execute on such a device." |
| 129 | ] |
| 130 | }, |
| 131 | { |
| 132 | "cell_type": "code", |
| 133 | "execution_count": null, |
| 134 | "id": "e2422917", |
| 135 | "metadata": {}, |
| 136 | "outputs": [], |
| 137 | "source": [ |
| 138 | "# Compile the code into a high-level program\n", |
| 139 | "program = compile(code.ParallelTeleport, 120)\n", |
| 140 | "\n", |
| 141 | "# Visualize a trace of running the program on the target machine,\n", |
| 142 | "# decomposing into native gates and scheduling parallel operations\n", |
| 143 | "device = NeutralAtomDevice()\n", |
| 144 | "device.show_trace(program)" |
| 145 | ] |
| 146 | }, |
| 147 | { |
| 148 | "cell_type": "markdown", |
| 149 | "id": "87b8e819", |
| 150 | "metadata": {}, |
| 151 | "source": [ |
| 152 | "Because the teleport algorithm uses only gates from the Clifford group, we can efficiently simulate a large number of shots of the same program with the Clifford simulator. We'll use a small helper function to interpret the results, which we expect to be all zeros. Without noise, \"Correct\" should be the only bar in the resulting Histogram." |
| 153 | ] |
| 154 | }, |
| 155 | { |
| 156 | "cell_type": "code", |
| 157 | "execution_count": null, |
| 158 | "id": "698849c0", |
| 159 | "metadata": {}, |
| 160 | "outputs": [], |
| 161 | "source": [ |
| 162 | "results = device.simulate(program, shots=1000, type=\"clifford\")\n", |
| 163 | "Histogram([\"Loss\" if any(r == Result.Loss for r in shot) else \"Flip\" if any(r == Result.One for r in shot) else \"Correct\" for shot in results])" |
| 164 | ] |
| 165 | }, |
| 166 | { |
| 167 | "cell_type": "markdown", |
| 168 | "id": "ce124bf0", |
| 169 | "metadata": {}, |
| 170 | "source": [ |
| 171 | "You can also perform this simulation with noise, shown here by configuring a 0.1% chance of loss on the SX gate. Now we'll see a mix of correct and incorrect results in the output of the simulation." |
| 172 | ] |
| 173 | }, |
| 174 | { |
| 175 | "cell_type": "code", |
| 176 | "execution_count": null, |
| 177 | "id": "774df3a7", |
| 178 | "metadata": {}, |
| 179 | "outputs": [], |
| 180 | "source": [ |
| 181 | "# Now with noise.\n", |
| 182 | "noise = NoiseConfig()\n", |
| 183 | "noise.sx.loss = 0.001\n", |
| 184 | "results = device.simulate(program, shots=1000, noise=noise, type=\"clifford\")\n", |
| 185 | "Histogram([\"Loss\" if any(r == Result.Loss for r in shot) else \"Flip\" if any(r == Result.One for r in shot) else \"Correct\" for shot in results])" |
| 186 | ] |
| 187 | } |
| 188 | ], |
| 189 | "metadata": { |
| 190 | "kernelspec": { |
| 191 | "display_name": "Python 3", |
| 192 | "language": "python", |
| 193 | "name": "python3" |
| 194 | }, |
| 195 | "language_info": { |
| 196 | "codemirror_mode": { |
| 197 | "name": "ipython", |
| 198 | "version": 3 |
| 199 | }, |
| 200 | "file_extension": ".py", |
| 201 | "mimetype": "text/x-python", |
| 202 | "name": "python", |
| 203 | "nbconvert_exporter": "python", |
| 204 | "pygments_lexer": "ipython3", |
| 205 | "version": "3.11.14" |
| 206 | } |
| 207 | }, |
| 208 | "nbformat": 4, |
| 209 | "nbformat_minor": 5 |
| 210 | } |
| 211 | |