{ "cells": [ { "cell_type": "markdown", "id": "797fe94d-93a3-4a7b-8d60-0706d5ab21d5", "metadata": {}, "source": [ "---\n", "title: \"Comparer les paramètres du transpilateur\"\n", "description: \"découvrez le pipeline de transcompilation tout au long du processus complet de création, de transcompilation et de soumission des circuits.\"\n", "---\n", "\n", "\n", "\n", "# Comparer les paramètres du transpilateur\n", "\n" ] }, { "cell_type": "markdown", "id": "d403684a-9dc5-433b-a788-789881878d6c", "metadata": { "tags": [ "version-info" ] }, "source": [ "{/*\n", " DO NOT EDIT THIS CELL!!!\n", " This cell's content is generated automatically by a script. Anything you add\n", " here will be removed next time the notebook is run. To add new content, create\n", " a new cell before or after this one.\n", " */}\n", "\n", "\n", " \n", " Le code présenté sur cette page a été développé en tenant compte des exigences suivantes.\n", " Nous vous recommandons d'utiliser ces versions ou des versions plus récentes.\n", "\n", " ```\n", " qiskit[all]~=2.5.1\n", " qiskit-ibm-runtime~=0.47.0\n", " ```\n", " \n", "\n", "\n" ] }, { "cell_type": "markdown", "id": "a6affcc2-72f4-4f06-8c4c-fc52715b0285", "metadata": {}, "source": [ "Les différents paramètres du transpileur permettent d'appliquer divers types d'optimisation au circuit, souvent au prix d'un allongement du temps de traitement classique. Ce guide vous accompagne tout au long du processus de création, de compilation et de soumission de circuits afin de vous permettre de tester les performances de différents paramètres.\n", "\n", "Il convient de noter qu'un même réglage peut améliorer les performances d'un circuit tout en nuisant à celles d'un autre. Veillez à vérifier les circuits transpilés obtenus avant de les exécuter sur du matériel réel.\n", "\n" ] }, { "cell_type": "markdown", "id": "39f9c961-c52d-46fc-a2aa-464462474b56", "metadata": {}, "source": [ "\n", "\n", "## Configurer et créer un circuit type\n", "\n" ] }, { "cell_type": "code", "execution_count": 1, "id": "790b4934-ae24-4e69-be9f-d82ae639a5e6", "metadata": {}, "outputs": [], "source": [ "# Create circuit to test transpiler on\n", "from qiskit import QuantumCircuit\n", "from qiskit.transpiler.preset_passmanagers import generate_preset_pass_manager\n", "from qiskit.circuit.library import grover_operator, DiagonalGate\n", "\n", "# Use Statevector object to calculate the ideal output\n", "from qiskit.quantum_info import Statevector\n", "from qiskit.visualization import plot_histogram\n", "from qiskit.transpiler import PassManager\n", "\n", "from qiskit.circuit.library import XGate\n", "from qiskit.quantum_info import hellinger_fidelity" ] }, { "cell_type": "markdown", "id": "fe0a4958-b406-4fe4-9415-38a772ad152c", "metadata": {}, "source": [ "Créez un petit circuit que le transcompilateur pourra essayer d'optimiser. Cet exemple crée un circuit qui exécute l'algorithme de Grover avec un oracle qui marque l'état `111`. Ensuite, simulez la distribution idéale (ce que vous vous attendriez à mesurer si vous exécutiez cela un nombre infini de fois sur un ordinateur quantique parfait) pour pouvoir la comparer ultérieurement.\n", "\n" ] }, { "cell_type": "code", "execution_count": 2, "id": "4ac958d4-b9b5-4939-a359-a9edca7ddb6a", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "\"Output" ] }, "execution_count": 2, "metadata": {}, "output_type": "execute_result" } ], "source": [ "oracle = DiagonalGate([1] * 7 + [-1])\n", "qc = QuantumCircuit(3)\n", "qc.h([0, 1, 2])\n", "qc = qc.compose(grover_operator(oracle))\n", "\n", "qc.draw(output=\"mpl\", style=\"iqp\")" ] }, { "cell_type": "code", "execution_count": 3, "id": "6313186e-bc40-432e-9ada-8594d6a26d55", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "\"Output" ] }, "execution_count": 3, "metadata": {}, "output_type": "execute_result" } ], "source": [ "ideal_distribution = Statevector.from_instruction(qc).probabilities_dict()\n", "\n", "plot_histogram(ideal_distribution)" ] }, { "cell_type": "markdown", "id": "964ca1e0-d1c9-40ed-bcf1-babc50f847ed", "metadata": {}, "source": [ "\n", "\n", "## Transpiler\n", "\n", "Ensuite, transcompilez les circuits pour le QPU. Vous comparerez les performances du transcompilateur avec `optimization_level` réglé sur `0` (le plus bas) par rapport à `3` (le plus élevé). Le niveau d'optimisation le plus bas effectue le strict minimum nécessaire pour faire fonctionner le circuit sur le dispositif; il mappe les qubits du circuit aux qubits du dispositif et ajoute des portes d'échange pour permettre toutes les opérations à deux qubits. Le niveau d'optimisation le plus élevé est beaucoup plus intelligent et utilise de nombreuses astuces pour réduire le nombre total de portes. Étant donné que les portes multi-qubits ont des taux d'erreur élevés et que les qubits se décohèrent avec le temps, les circuits plus courts devraient donner de meilleurs résultats.\n", "\n", "\n", " Cet exemple utilise le matériel d' IBM Quantum®, mais vous pouvez l'essayer sur n'importe quel QPU compatible avec Qiskit. Vos résultats peuvent varier.\n", "\n", "\n", "La cellule suivante transcompile `qc` pour les deux valeurs de `optimization_level`, affiche le nombre de portes à deux qubits et ajoute les circuits transcompilés à une liste. Certains algorithmes du transcompilateur sont aléatoires, il définit donc une graine pour assurer la reproductibilité.\n", "\n" ] }, { "cell_type": "code", "execution_count": 4, "id": "61181ac0-3f89-417f-a31e-9430f63e670b", "metadata": {}, "outputs": [], "source": [ "# Use IBM Quantum Compute Service to run jobs on hardware\n", "from qiskit_ibm_runtime import (\n", " QiskitRuntimeService,\n", " SamplerV2 as Sampler,\n", ")" ] }, { "cell_type": "code", "execution_count": 5, "id": "c3062a60-1cdc-46e7-8eb3-efc62a1396bd", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "'ibm_fez'" ] }, "execution_count": 5, "metadata": {}, "output_type": "execute_result" } ], "source": [ "# Select the backend with the fewest number of jobs in the queue\n", "service = QiskitRuntimeService()\n", "backend = service.least_busy(\n", " operational=True, simulator=False, min_num_qubits=127\n", ")\n", "backend.name" ] }, { "cell_type": "code", "execution_count": 6, "id": "2a3ebe8c-e47d-4440-b004-f47f6af826f0", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Two-qubit gates (optimization_level=0): 21\n", "Two-qubit gates (optimization_level=3): 12\n" ] } ], "source": [ "# Need to add measurements to the circuit\n", "qc.measure_all()\n", "\n", "# Find the correct two-qubit gate\n", "twoQ_gates = set([\"ecr\", \"cz\", \"cx\"])\n", "for gate in backend.basis_gates:\n", " if gate in twoQ_gates:\n", " twoQ_gate = gate\n", "\n", "circuits = []\n", "for optimization_level in [0, 3]:\n", " pm = generate_preset_pass_manager(\n", " optimization_level, backend=backend, seed_transpiler=0\n", " )\n", " t_qc = pm.run(qc)\n", " print(\n", " f\"Two-qubit gates (optimization_level={optimization_level}): \",\n", " t_qc.count_ops()[twoQ_gate],\n", " )\n", " circuits.append(t_qc)" ] }, { "cell_type": "markdown", "id": "99928d6b-a7e7-40c9-b59c-104fc430b57c", "metadata": {}, "source": [ "Étant donné que les CNOT ont généralement un taux d'erreur élevé, le circuit transpilé avec `optimization_level=3` devrait offrir de bien meilleures performances.\n", "\n", "Une autre façon d'améliorer les performances consiste à recourir [au découplage dynamique](/docs/api/qiskit/qiskit.transpiler.passes.PadDynamicalDecoupling), en appliquant une séquence de portes à des qubits au repos. Cela permet d'éliminer certaines interactions indésirables avec l'environnement. La cellule suivante ajoute un découplage dynamique au circuit transpilé avec `optimization_level=3` et l'ajoute à la liste.\n", "\n" ] }, { "cell_type": "code", "execution_count": 7, "id": "b20ebca3-4adb-4a95-9f6a-bb4cbd836daf", "metadata": {}, "outputs": [], "source": [ "from qiskit_ibm_runtime.transpiler.passes.scheduling import (\n", " ASAPScheduleAnalysis,\n", " PadDynamicalDecoupling,\n", ")\n", "\n", "# Get gate durations so the transpiler knows how long each operation takes\n", "durations = backend.target.durations()\n", "\n", "# This is the sequence we'll apply to idling qubits\n", "dd_sequence = [XGate(), XGate()]\n", "\n", "# Run scheduling and dynamic decoupling passes on circuit\n", "pm = PassManager(\n", " [\n", " ASAPScheduleAnalysis(durations),\n", " PadDynamicalDecoupling(durations, dd_sequence),\n", " ]\n", ")\n", "circ_dd = pm.run(circuits[1])\n", "\n", "# Add this new circuit to our list\n", "circuits.append(circ_dd)" ] }, { "cell_type": "code", "execution_count": 8, "id": "c1c91fbd-acfe-413e-a6c9-ad97f4dd5543", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "\"Output" ] }, "execution_count": 8, "metadata": {}, "output_type": "execute_result" } ], "source": [ "circ_dd.draw(output=\"mpl\", style=\"iqp\", idle_wires=False)" ] }, { "cell_type": "markdown", "id": "5bcc75c7-8af0-4862-8e3c-ec4d06aed1f1", "metadata": {}, "source": [ "\n", "\n", "## Effectuez le circuit\n", "\n", "À ce stade, vous disposez d'une liste de circuits transpilés avec différents paramètres. Ensuite, exécutez ces circuits à l'aide de la primitive Sampler et enregistrez les résultats dans `result`.\n", "\n" ] }, { "cell_type": "code", "execution_count": 9, "id": "c1b36384-fd9b-4e24-a399-32d35fc6fa5b", "metadata": {}, "outputs": [], "source": [ "sampler = Sampler(backend)\n", "job = sampler.run(\n", " [(circuit) for circuit in circuits], # sample all three circuits\n", " shots=8000,\n", ")\n", "result = job.result()" ] }, { "cell_type": "markdown", "id": "c85b4da4-592f-45a6-87a2-a8f2d3415576", "metadata": {}, "source": [ "\n", "\n", "## Afficher les résultats\n", "\n", "Enfin, tracez les résultats des essais sur l'appareil en fonction de la distribution idéale. On constate que les résultats obtenus avec `optimization_level=3` se rapprochent davantage de la distribution idéale grâce au nombre réduit de grilles, et que `optimization_level=3 + dd` s'en rapproche encore davantage grâce au découplage dynamique.\n", "\n" ] }, { "cell_type": "code", "execution_count": 10, "id": "9e86132d-a8b2-40db-af42-53042dfa108b", "metadata": {}, "outputs": [ { "data": { "text/plain": [ "\"Output" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "binary_prob = [\n", " {\n", " k: v / res.data.meas.num_shots\n", " for k, v in res.data.meas.get_counts().items()\n", " }\n", " for res in result\n", "]\n", "plot_histogram(\n", " binary_prob + [ideal_distribution],\n", " bar_labels=False,\n", " legend=[\n", " \"optimization_level=0\",\n", " \"optimization_level=3\",\n", " \"optimization_level=3 + dd\",\n", " \"ideal distribution\",\n", " ],\n", ")" ] }, { "cell_type": "markdown", "id": "47a9eec8-7b31-4b2d-a291-559ddfd7a36b", "metadata": {}, "source": [ "Vous pouvez le vérifier en calculant la [fidélité de Hellinger](/docs/api/qiskit/quantum_info) entre chaque ensemble de résultats et la distribution idéale (plus le chiffre est élevé, mieux c'est, et 1 correspond à une fidélité parfaite).\n", "\n" ] }, { "cell_type": "code", "execution_count": 11, "id": "d2b5e797-176b-48b9-ac2b-ba73abe9300f", "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "0.774\n", "0.978\n", "0.979\n" ] } ], "source": [ "for prob in binary_prob:\n", " print(f\"{hellinger_fidelity(prob, ideal_distribution):.3f}\")" ] }, { "cell_type": "markdown", "id": "1b5b7bb9-eedb-45eb-a4cf-9b7708cbbb3e", "metadata": {}, "source": [ "\n", "\n", "## Etapes suivantes\n", "\n", "\n", " * Découvrez quelques ressources avancées sur la transpilation, telles que :\n", "\n", " * [Écrire une passe de transpilation personnalisée](/docs/guides/custom-transpiler-pass)\n", " * [Créer et transcompiler pour des backends personnalisés](/docs/guides/custom-backend)\n", " * [Installer et utiliser des plugins de transcompilation](/docs/guides/transpiler-plugins)\n", "\n", " * Parcourez les [tutoriels](/docs/tutorials) disponibles.\n", "\n", "\n" ] }, { "cell_type": "markdown", "metadata": {}, "id": "a1b8767d", "source": "© IBM Corp., 2017-2026" } ], "metadata": { "kernelspec": { "display_name": "Python 3", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3" } }, "nbformat": 4, "nbformat_minor": 4 }