แก้ไข

แชร์ผ่าน


Understanding the result data of the Resource Estimator

Learn how to interpret and retrieve the output parameters and diagrams of the Resource Estimator. This article explains how to programmatically access the results of the Resource Estimator in Jupyter Notebooks. If you run the Resource Estimator in Visual Studio Code from the command palette, the following commands doesn't apply.

Prerequisites

Output parameters

The output data of the Resource Estimator is a report that is printed in the console and can be accessed programmatically. For example, the following code snippet shows how to access the resource estimation parameters.

result['jobParams']

The following output data constitutes the possible entries that can be access programmatically.

Top-level output parameter Data type Description
status string The status of the job, it's always Succeeded.
jobParams dictionary The target parameters of the job that are passed as input.
physicalCounts dictionary The physical resource estimates. For more information, see Physical counts.
physicalCountsFormatted dictionary The physical resource estimates formatted for display in report data. For more information, see Physical counts formatted.
logicalQubit dictionary The logical qubit properties. For more information, see Logical qubit.
tfactory dictionary The T factory properties.
logicalCounts dictionary The pre-layout logical resource estimates. For more information, see Logical counts.
reportData dictionary Generation data for resource estimation report.

Physical counts

The physicalCounts dictionary contains the following entries:

Output parameter Data type Description
physicalQubits number The total number of physical qubits.
runtime number The total runtime to execute the algorithm in nanoseconds.
rqops number The number of reliable quantum operations per second (QOPS).
breakdown dictionary Breakdown of estimates. For more information, see Physical counts breakdown.

Physical counts breakdown

The breakdown dictionary of physicalCounts contains the following entries:

Output parameter Data type Description
algorithmicLogicalQubits number The logical qubits required for running the algorithm and do not include resources for T factories.
algorithmicLogicalDepth number The logical cycles required for running the algorithm and do not include resources for T factories.
logicalDepth number The possibly adjusted number of cycles that is computed whenever the T factory execution time is faster then algorithm execution.
numTstates number The number of T states consumed by the algorithm.
clockFrequency number The number of logical cycles per second.
numTfactories number The number of T factories (assuming uniform T factory design).
numTfactoryRuns number The number of how often all parallel T factories should run.
physicalQubitsForTfactories number The number of physical qubits for all T factories.
physicalQubitsForAlgorithm number The number of physical qubits for algorithm layout.
requiredLogicalQubitErrorRate number The required logical error rate.
requiredLogicalTstateErrorRate number The required logical T state error rate.
numTsPerRotation number The number of T gates per rotation.
cliffordErrorRate number The Clifford error rate based on the qubit parameters.

Physical counts formatted

The physicalCountsFormatted dictionary contains the following entries:

Output parameter Data type Description
runtime string Total runtime as human friendly string.
rqops string The number of reliable quantum operations per second (QOPS) formatted with metric suffix.
physicalQubits string Total number of physical qubits with metric suffix.
algorithmicLogicalQubits string Algorithmic logical qubits with metric suffix.
algorithmicLogicalDepth string Algorithmic logical depth with metric suffix.
logicalDepth string Possibly adjusted algorithmic logical depth with metric suffix.
numTstates string Number of T states with metric suffix.
numTfactories string Number of T factory copies with metric suffix.
numTfactoryRuns string Number of T factory runs with metric suffix.
physicalQubitsForAlgorithm string Number of physical qubits for algorithm with metric suffix.
physicalQubitsForTfactories string Number of physical qubits for T factories with metric suffix.
physicalQubitsForTfactoriesPercentage string The number of physical qubits for all T factories in percentage to total.
requiredLogicalQubitErrorRate string Truncated required logical qubit error rate.
requiredLogicalTstateErrorRate string Truncated required T state error rate.
physicalQubitsPerLogicalQubit string Number of physical qubits per logical qubit with metric suffix.
logicalCycleTime string The logical cycle time of a logical qubit as human friendly string.
clockFrequency string The number of logical cycles per second as a human friendly string.
logicalErrorRate string Truncated logical error rate.
tfactoryPhysicalQubits string Number of physical qubits in T factory with metric suffix (or message that there is no T factory).
tfactoryRuntime string The runtime of a single T factory as human friendly string (or message that there is no T factory).
numInputTstates string The number of input T states (or message that there is no T factory).
numUnitsPerRound string The number of units per distillation round, comma separated in a string (or message that there is no T factory).
unitNamePerRound string The unit names of each distillation round, comma separated in a string (or message that there is no T factory).
codeDistancePerRound string The code distances per distillation round, comma separated in a string (or message that there is no T factory).
physicalQubitsPerRound string The number of physical qubits per distillation round, comma separated in a string (or message that there is no T factory).
tfactoryRuntimePerRound string The runtime of each distillation round, displayed as comma separated human friendly strings (or message that there is no T factory).
tstateLogicalErrorRate string Truncated logical T state error rate (or message that there is no T factory).
logicalCountsNumQubits string Number of qubits (pre-layout) with metric suffix.
logicalCountsTCount string Number of T gates (pre-layout) with metric suffix.
logicalCountsRotationCount string Number of rotation gates (pre-layout) with metric suffix.
logicalCountsRotationDepth string Rotation depth (pre-layout) with metric suffix.
logicalCountsCczCount string Number of CCZ gates (pre-layout) with metric suffix.
logicalCountsCcixCount string Number of CCiX gates (pre-layout) with metric suffix.
logicalCountsMeasurementCount string Number of single-qubit measurements (pre-layout) with metric suffix.
errorBudget string Truncated total error budget.
errorBudgetLogical string Truncated error budget for logical error.
errorBudgetTstates string Truncated error budget for faulty T state distillation.
errorBudgetRotations string Truncated error budget for faulty rotation synthesis.
numTsPerRotation string Formatted number of Ts per rotation (might be None).

Logical qubit

The logicalQubit dictionary contains the following entries:

Output parameter Data type Description
codeDistance number The computed code distance for the logical qubit.
physicalQubits number The number of physical qubits for each logical qubit.
logicalCycleTime number The time to execute one logical operation.
logicalErrorRate number The logical error rate of the logical qubit.

Logical counts

The logicalCounts dictionary contains the following entries:

Output parameter Data type Description
numQubits number Pre-layout number of qubits.
tCount number Pre-layout number of T gates.
rotationCount number Pre-layout number of rotation gates.
rotationDepth number Pre-layout rotation depth.
cczCount number Pre-layout number of CCZ gates.
ccixCount number Pre-layout number of CCiX gates.
measurementCount number Pre-layout number of single-qubit measurements.

Tip

If you want to use a pre-calculated set of logical counts for a resource estimation job, you can use the LogicalCounts Python operation to pass the known estimates to the Resource Estimator. For more information, see How to use known estimates with the Resource Estimator.

Space diagram

The overall physical resource estimation consists of total number of physical qubits used for both the algorithm and T factory copies. You can inspect the distribution between these two using the space diagram.

The space diagram shows the proportion of the physical qubits used for the algorithm and the T factories. Note that the number of T factory copies contributes to the number of physical qubits for T factories.

In Jupyter Notebook, you can access the space diagram using the SpaceChart widget from the qsharp-widgets package.

import qsharp

from qsharp_widgets import SpaceChart
SpaceChart(result)

Pie diagram showing the distribution of total physical qubits between algorithm qubits and T factory qubits. There's a table with the breakdown of number of T factory copies and number of physical qubits per T factory.

When running multiple configurations of target parameters with the Pareto frontier estimation, you can plot the space diagram for a specific solution of the. For example, the following code shows how to plot the space diagram for the first configuration of parameters and the third shortest runtime.

SpaceChart(result[0], 2) # First (estimate index=0) run and third (point index=2) shortest runtime

Space-time diagram

In quantum computing, there's a tradeoff between the number of physical qubits and the runtime of the algorithm. You could consider allocation of as many physical qubits as possible to reduce the runtime of the algorithm. However, the number of physical qubits is limited by the number of physical qubits available in the quantum hardware. Understanding the tradeoff between runtime and system scale is one of the more important aspects of resource estimation.

When estimating the resources of an algorithm, you can use the space-time diagram to visualize the tradeoffs between the number of physical qubits and the runtime of the algorithm.

Note

To see multiple optimal combinations in the space-time diagram, you need to set the estimation type to Pareto frontier estimation. If you run the Resource Estimator in Visual Studio Code using the Q#: Calculate Resource Estimates option, the Pareto frontier estimation is enabled by default.

The space-time diagram allows you to find the optimal combination of {number of qubits, runtime} pairs that satisfy the constraints of the quantum hardware. The diagram shows the number of physical qubits and the runtime of the algorithm for each {number of qubits, runtime} pair.

To run the space-time diagram in Jupyter Notebook, you can use the EstimatesOverview widget from the qsharp-widgets package.

import qsharp

from qsharp_widgets import EstimatesOverview

EstimatesOverview(result, colors=["#1f77b4", "#ff7f0e"], runNames=["e4 Surface Code", "e6 Floquet Code"])

Screenshot showing the qubit-time diagram of the Resource Estimator.

Tip

To see the estimation details, you can hover over each point in the diagram.

The space-time diagram is specially useful when comparing multiple configurations of target parameters for the same algorithm.

Screenshot showing the space-time diagram and the table of results when running multiple configurations of parameter in the Resource Estimator.

Note

If you run into any issue while working with the Resource Estimator, check out the Troubleshooting page, or contact AzureQuantumInfo@microsoft.com.

Next steps