Amazon Braket terms and concepts - Amazon Braket

Amazon Braket terms and concepts

Tip

Learn the foundations of quantum computing with AWS! Enroll in the Amazon Braket Digital Learning Plan and earn your own Digital badge after completing a series of learning courses and a digital assessment.

The following terms and concepts are used in Braket:

Analog Hamiltonian Simulation

Analog Hamiltonian Simulation (AHS) is a distinct quantum computing paradigm for direct simulation of time-dependent quantum dynamics of many-body systems. In AHS, users directly specify a time-dependent Hamiltonian and the quantum computer is tuned in such a way that it directly emulates the continuous time evolution under this Hamiltonian. AHS devices are typically special-purpose devices and not universal quantum computers like gate-based devices. They are limited to a class of Hamiltonians they can simulate. However, since these Hamiltonians are naturally implemented on the device, AHS does not suffer from the overhead required to formulate algorithms as circuits and implement gate operations.

Braket

We named the Braket service after the bra-ket notation, a standard notation in quantum mechanics. It was introduced by Paul Dirac in 1939 to describe the state of quantum systems, and it is also known as the Dirac notation.

Braket Direct

With Braket Direct, you can reserve dedicated access to different quantum devices of your choice, connect with quantum computing specialists to receive guidance for your workload, and gain early access to next-generation capabilities, such as new quantum devices with limited availability.

Braket hybrid job

Amazon Braket has a feature called Amazon Braket Hybrid Jobs that provides fully managed executions of hybrid algorithms. A Braket hybrid job consists of three components:

  1. The definition of your algorithm, which can be provided as a script, Python module, or Docker container.

  2. The hybrid job instance, based on Amazon EC2, on which to run your algorithm. The default is an ml.m5.xlarge instance.

  3. The quantum device on which to run the quantum tasks that are part of your algorithm. A single hybrid job typically contains a collection of many quantum tasks.

Device

In Amazon Braket, a device is a backend that can run quantum tasks. A device can be a QPU or a quantum circuit simulator. To learn more, see Amazon Braket supported devices.

Gate-based quantum computing

In gate-based quantum computing (QC), also called circuit-based QC, computations are broken down into elementary operations (gates). Certain sets of gates are universal, meaning that every computation can be expressed as a finite sequence of those gates. Gates are the building blocks of quantum circuits and are analogous to the logic gates of classical digital circuits.

Hamiltonian

The quantum dynamics of a physical system are determined by its Hamiltonian, which encodes all information about the interactions between constituents of the system and the effects of exogenous driving forces. The Hamiltonian of an N-qubit system is commonly represented as a 2N by 2N matrix of complex numbers on classical machines. By running an Analog Hamiltonian Simulation on a quantum device, you can avoid these exponential resource requirements.

Pulse

A pulse is a transient physical signal transmitted to the qubits. It is described by a waveform played in a frame that serves as a support for the carrier signal and is bound to the hardware channel or port. Customers can design their own pulses by providing the analog envelope that modulates the high-frequency sinusoidal carrier signal. The frame is uniquely described by a frequency and a phase that are often chosen to be on resonance with the energy separation between the energy levels for |0⟩ and |1⟩ of the qubit. Gates are thus enacted as pulses with a predetermined shape and calibrated parameters such as its amplitude, frequency and duration. Use cases that are not covered by template waveforms will be enabled through custom waveforms which will be specified at the single sample resolution by providing a list of values separated by a fixed, physical cycle-time.

Quantum circuit

A quantum circuit is the instruction set that defines a computation on a gate-based quantum computer. A quantum circuit is a sequence of quantum gates, which are reversible transformations on a qubit register, together with measurement instructions.

Quantum circuit simulator

A quantum circuit simulator is a computer program that runs on classical computers and calculates the measurement outcomes of a quantum circuit. For general circuits, the resource requirements of a quantum simulation grow exponentially with the number of qubits to simulate. Braket provides access to both managed (accessed through the Braket API) and local (part of the Amazon Braket SDK) quantum circuit simulators.

Quantum computer

A quantum computer is a physical device that uses quantum-mechanical phenomena, such as superposition and entanglement, to perform computations. There are different paradigms to quantum computing (QC), such as gate-based QC.

Quantum processing unit (QPU)

A QPU is a physical quantum computing device that can run on a quantum task. QPUs can be based on different QC paradigms, such as gate-based QC. To learn more, see Amazon Braket supported devices.

QPU native gates

QPU native gates can be directly mapped to control pulses by the QPU control system. Native gates can be run on the QPU device without further compilation. Subset of QPU supported gates. You can find the native gates of a device on the Devices page in the Amazon Braket console and through the Braket SDK.

QPU supported gates

QPU supported gates are the gates accepted by the QPU device. These gates might not be able to directly run on the QPU, meaning that they might need to be decomposed into native gates. You can find the supported gates of a device on the Devices page in the Amazon Braket console and through the Amazon Braket SDK.

Quantum task

In Braket, a quantum task is the atomic request to a device. For gate-based QC devices, this includes the quantum circuit (including the measurement instructions and number of shots) and other request metadata. You can create quantum tasks through the Amazon Braket SDK or by using the CreateQuantumTask API operation directly. After you create a quantum task, it will be queued until the requested device becomes available. You can view your quantum tasks on the Quantum Tasks page of the Amazon Braket console or by using the GetQuantumTask or SearchQuantumTasks API operations.

Qubit

The basic unit of information in a quantum computer is called a qubit (quantum bit), much like a bit in classical computing. A qubit is a two-level quantum system that can be realized by different physical implementations, such as superconducting circuits or individual ions and atoms. Other qubit types are based on photons, electronic or nuclear spins, or more exotic quantum systems.

Queue depth

Queue depth refers to the number of quantum tasks and hybrid jobs queued for a particular device. A device’s quantum task and hybrid job queue count are accessible through the Braket Software Development Kit (SDK) or Amazon Braket Management Console.

  1. Task queue depth refers to the total number of quantum tasks currently waiting to run in normal priority.

  2. Priority task queue depth refers to the total number of submitted quantum tasks waiting to run through Amazon Braket Hybrid Jobs. These tasks get priority over standalone tasks once a hybrid job starts.

  3. Hybrid jobs queue depth refers to the total number of hybrid jobs currently queued on a device. Quantum tasks submitted as part of a hybrid job have priority, and are aggregated in the Priority Task Queue.

Queue position

Queue position refers to the current position of your quantum task or hybrid job within a respective device queue. It can be obtained for quantum tasks or hybrid jobs through the Braket Software Development Kit (SDK) or Amazon Braket Management Console.

Shots

Since quantum computing is inherently probabilistic, any circuit needs to be evaluated multiple times to get an accurate result. A single circuit execution and measurement is called a shot. The number of shots (repeated executions) for a circuit is chosen based on the desired accuracy for the result.

AWS terminology and tips for Amazon Braket

IAM policies

An IAM policy is a document that allows or denies permissions to AWS services and resources. IAM policies enable you to customize users' levels of access to resources. For example, you can allow users access to all of the Amazon S3 buckets within your AWS account, or only a specific bucket.

  • Best practice: Follow the security principle of least privilege when granting permissions. By following this principle, you help to prevent users or roles from having more permissions than needed to perform their quantum tasks. For example, if an employee needs access to only a specific bucket, specify the bucket in the IAM policy instead of granting the employee access to all of the buckets in your AWS account.

IAM roles

An IAM role is an identity that you can assume to gain temporary access to permissions. Before a user, application, or service can assume an IAM role, they must be granted permissions to switch to the role. When someone assumes an IAM role, they abandon all previous permissions that they had under a previous role and assume the permissions of the new role.

  • Best practice: IAM roles are ideal for situations in which access to services or resources needs to be granted temporarily, instead of long-term.

Amazon S3 bucket

Amazon Simple Storage Service (Amazon S3) is an AWS service that lets you store data as objects in buckets. Amazon S3 buckets offer unlimited storage space. The maximum size for an object in an Amazon S3 bucket is 5 TB. You can upload any type of file data to an Amazon S3 bucket, such as images, videos, text files, backup files, media files for a website, archived documents, and your Braket quantum task results.

  • Best practice: You can set permissions to control access to your S3 bucket. For more information, see Bucket policies in the Amazon S3 documentation.