feat(qvol): First version of adding Quantum Volume to Recirq.

recirq/benchmarks/Quantum-Volume.ipynb

@@ -0,0 +1,94 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Running the Quantum Volume Algorithm\n",
+    "This example runs the quantum volume algorithm of https://arxiv.org/pdf/1811.12926.pdf. In general, the algorithm will generate a model circuit, classically compute its Heavy Output Group, then run a sampler to evaluate how often it generates Heavy results.\n",
+    "\n",
+    "TODO: Flesh out this notebook per https://github.com/quantumlib/ReCirq/issues/28.\n",
+    "\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# This cell sets up the parameters for the quantum volume algorithm.\n",
+    "# Feel free to mess with these!\n",
+    "from recirq.benchmarks.experiments.quantum_volume import QuantumVolumeTask, run_quantum_volume\n",
+    "import datetime\n",
+    "\n",
+    "dataset_id = datetime.datetime.now().isoformat(timespec='minutes')\n",
+    "dc_task = QuantumVolumeTask(\n",
+    "    dataset_id=dataset_id,\n",
+    "    device_name='Syc23-simulator',\n",
+    "    n_shots=10_000,\n",
+    "    n_circuits=1,\n",
+    "    n_qubits=4,\n",
+    "    depth=4,\n",
+    "    readout_error_correction=True,\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# Actually run the task and collect the data for analysis.\n",
+    "from recirq.benchmarks.experiments.quantum_volume import EXPERIMENT_NAME, DEFAULT_BASE_DIR, QuantumVolumeTask, run_quantum_volume\n",
+    "\n",
+    "run_quantum_volume(dc_task)\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import recirq\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "\n",
+    "records = []\n",
+    "# Load all data, do some light processing\n",
+    "for record in recirq.iterload_records(dataset_id=dataset_id, base_dir=DEFAULT_BASE_DIR):\n",
+    "    # Expand task dataclass into columns\n",
+    "    recirq.flatten_dataclass_into_record(record, 'task')\n",
+    "\n",
+    "    records.append(record)\n",
+    "    \n",
+    "df = pd.DataFrame(records)\n",
+    "print(len(df))\n",
+    "df.head()"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "ReCirq",
+   "language": "python",
+   "name": "recirq"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.7.7"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 4
+}

recirq/benchmarks/__init__.py

@@ -0,0 +1,13 @@
+# Copyright 2020 Google
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.

recirq/benchmarks/experiments/__init__.py

@@ -0,0 +1,13 @@
+# Copyright 2020 Google
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.

recirq/benchmarks/experiments/quantum_volume.py

@@ -0,0 +1,129 @@
+# Copyright 2020 Google
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+from typing import Optional, List
+
+import numpy as np
+
+import cirq
+from cirq.contrib.quantum_volume import quantum_volume
+
+import recirq
+
+
+@recirq.json_serializable_dataclass(namespace='recirq.quantum_volume',
+                                    registry=recirq.Registry,
+                                    frozen=True)
+class QuantumVolumeTask:
+    """Script to run the quantum volume algorithm on a device, per the benchmark defined by IBM in
+    https://arxiv.org/abs/1811.12926.
+
+    This will run the Quantum Volume benchmark on the Sycamore device on the given depths (every combination of
+    depths 2 through 6 by default) with 100 repetitions each depth. You can configure these parameters by editing
+    this script directly. All the business logic is deferred to a Quantum Volume back in public Cirq - this file
+    exists primarily to pull in devices that we are not yet ready to release.
+
+    Attributes:
+        dataset_id: A unique identifier for this dataset.
+        device_name: The device to run on, by name.
+        num_qubits: The number of qubits in the generated circuits.
+        qubits: All of the possible qubits that the algorithm can run on. If empty, 
+            we will use any qubits on the device.
+        n_shots: The number of repetitions for each circuit.
+        n_circuits: The number of circuits to run the algorithm.
+        depth: The number of layers to generate in the circuit.
+    """
+    dataset_id: str
+    device_name: str
+    n_qubits: int
+    n_shots: int
+    n_circuits: int
+    depth: int
+    readout_error_correction: bool
+    qubits: Optional[List[cirq.GridQubit]] = None
+
+    @property
+    def fn(self):
+        n_shots = _abbrev_n_shots(self.n_shots)
+        qubits = _abbrev_grid_qubits(self.qubits) + '/'
+
+        return (f'{self.dataset_id}/'
+                f'{self.device_name}/'
+                f'{qubits}/'
+                f'{self.n_circuits}_{self.depth}_{self.n_qubits}_{n_shots}')
+
+
+# Define the following helper functions to make nicer `fn` keys
+# for the tasks:
+
+
+def _abbrev_n_shots(n_shots: int) -> str:
+    """Shorter n_shots component of a filename"""
+    if n_shots % 1000 == 0:
+        return f'{n_shots // 1000}k'
+    return str(n_shots)
+
+
+def _abbrev_grid_qubits(qubits: Optional[List[cirq.GridQubit]]) -> str:
+    """Formatted a list of grid qubits component of a filename"""
+    if not qubits:
+        return ''
+    return '-'.join([f'{qubit.row}_{qubit.col}' for qubit in qubits])
+
+
+EXPERIMENT_NAME = 'quantum-volume'
+DEFAULT_BASE_DIR = os.path.expanduser(f'~/cirq-results/{EXPERIMENT_NAME}')
+
+
+def run_quantum_volume(task: QuantumVolumeTask, base_dir=None):
+    """Execute a :py:class:`QuantumVolumeTask` task."""
+    if base_dir is None:
+        base_dir = DEFAULT_BASE_DIR
+
+    if recirq.exists(task, base_dir=base_dir):
+        print(f"{task} already exists. Skipping.")
+        return
+
+    sampler = recirq.get_sampler_by_name(device_name=task.device_name)
+    device = recirq.get_device_obj_by_name(device_name=task.device_name)
+    device_or_qubits = task.qubits if task.qubits else device
+
+    # Run the jobs
+    print("Collecting data", flush=True)
+    results = quantum_volume.calculate_quantum_volume(
+        num_qubits=task.n_qubits,
+        depth=task.depth,
+        num_circuits=task.n_circuits,
+        device_or_qubits=device_or_qubits,
+        samplers=[sampler],
+        repetitions=task.n_shots,
+        random_state=np.random.RandomState(
+            int(hash(task.dataset_id)) % (2**32 - 1)),
+        compiler=lambda c: cirq.google.optimized_for_sycamore(
+            c,
+            new_device=device,
+            optimizer_type='sycamore',
+            tabulation_resolution=0.008),
+        add_readout_error_correction=task.readout_error_correction)
+    # Save the results. NB: The circuits in the results have Circuits that contain a SerializableDevice, which at the
+    # time of this writing is not JSON serializable. Because we don't actually need the device to be in the circuits,
+    # we simply set them to None for now.
+    for result in results:
+        # We can't actually set the device property to None, so set _device directly.
+        result.compiled_circuit._device = None
+        result.model_circuit._device = None
+    recirq.save(task=task, data={
+        'results': results,
+    }, base_dir=base_dir)

recirq/benchmarks/experiments/run-quantum-volume.py

@@ -0,0 +1,48 @@
+# Copyright 2020 Google
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     https://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from recirq.benchmarks.experiments.quantum_volume import QuantumVolumeTask, run_quantum_volume
+import datetime
+
+
+def main():
+    """Main driver script entry point.
+
+    This function contains configuration options and you will likely need
+    to edit it to suit your needs. Of particular note, please make sure
+    `dataset_id` and `device_name`
+    are set how you want them. You may also want to change the values in
+    the list comprehension to set the qubits.
+    """
+    # Uncomment below for an auto-generated unique dataset_id
+    # dataset_id = datetime.datetime.now().isoformat(timespec='minutes')
+    dataset_id = '2020-05-21'
+    data_collection_tasks = [
+        QuantumVolumeTask(
+            dataset_id=dataset_id,
+            device_name='Syc23-simulator',
+            n_shots=10_000,
+            n_circuits=1,
+            n_qubits=4,
+            depth=4,
+            readout_error_correction=True,
+        )
+    ]
+
+    for dc_task in data_collection_tasks:
+        run_quantum_volume(dc_task)
+
+
+if __name__ == '__main__':
+    main()