Trait BatchExecution 

pub trait BatchExecution {
    // Provided methods
    fn sample(
        &self,
        _gates: &[GateInstruction],
        _qubits_to_sample: &[usize],
        _shots: usize,
    ) -> Result<SampleData, KetError> { ... }
    fn exp_value(
        &self,
        _gates: &[GateInstruction],
        _hamiltonian_list: &[Hamiltonian],
    ) -> Result<Vec<f64>, KetError> { ... }
    fn sample_native(
        &self,
        _gates: &[NativeGate],
        _qubits_to_sample: &[usize],
        _shots: usize,
    ) -> Result<SampleData, KetError> { ... }
    fn exp_value_native(
        &self,
        _gates: &[NativeGate],
        _hamiltonian_list: &[Hamiltonian],
    ) -> Result<Vec<f64>, KetError> { ... }
    fn gradient(
        &self,
        _gates: &[GateInstruction],
        _hamiltonian: &Hamiltonian,
    ) -> Result<(f64, Vec<f64>), KetError> { ... }
}

A backend that receives a fully-compiled gate sequence and returns aggregate results without mid-circuit feedback.

There are two families of methods: IR-level (sample, exp_value, gradient) which receive GateInstruction slices directly from the Libket IR, and native-level (sample_native, exp_value_native) which receive sequences of NativeGate tuples already translated through a NativeGateSet.

All methods have a default implementation that returns an error, so implementors only need to override the methods they support.

Provided Methods

fn sample( &self, _gates: &[GateInstruction], _qubits_to_sample: &[usize], _shots: usize, ) -> Result<SampleData, KetError>

Executes gates and samples the state of qubits_to_sample over shots shots, receiving gates in the Libket IR format.

Implement this method when the backend can accept GateInstruction slices directly (i.e., no NativeGateSet translation is configured).

Errors

Returns KetError::GateUnsupported by default. Implementations should return KetError::ExecutionFailed for backend faults or KetError::ShotCountInvalid for an out-of-range shot count.

fn exp_value( &self, _gates: &[GateInstruction], _hamiltonian_list: &[Hamiltonian], ) -> Result<Vec<f64>, KetError>

Executes gates and computes the expectation value of each Hamiltonian in hamiltonian_list, receiving gates in the Libket IR format.

Implement this method when the backend can accept GateInstruction slices directly (i.e., no NativeGateSet translation is configured).

Errors

Returns KetError::GateUnsupported by default. Implementations should return KetError::ExecutionFailed for backend faults.

fn sample_native( &self, _gates: &[NativeGate], _qubits_to_sample: &[usize], _shots: usize, ) -> Result<SampleData, KetError>

Executes gates and samples the state of qubits_to_sample over shots shots, receiving pre-translated native gates.

Called when a NativeGateSet translation layer is configured. The gates slice contains hardware-specific instructions produced by NativeGateSet::translate or NativeGateSet::cnot.

Errors

Returns KetError::GateUnsupported by default. Implementations should return KetError::ExecutionFailed for backend faults or KetError::ShotCountInvalid for an out-of-range shot count.

fn exp_value_native( &self, _gates: &[NativeGate], _hamiltonian_list: &[Hamiltonian], ) -> Result<Vec<f64>, KetError>

Executes gates and computes the expectation value of each Hamiltonian in hamiltonian_list, receiving pre-translated native gates.

Called when a NativeGateSet translation layer is configured. The gates slice contains hardware-specific instructions produced by NativeGateSet::translate or NativeGateSet::cnot.

Errors

Returns KetError::NativeGateUnsupported by default. Implementations should return KetError::ExecutionFailed for backend faults.

fn gradient( &self, _gates: &[GateInstruction], _hamiltonian: &Hamiltonian, ) -> Result<(f64, Vec<f64>), KetError>

Computes the expectation value and its gradient with respect to all circuit parameters in a single backend call.

Returns (expectation_value, gradient_vector) where gradient_vector[i] is ∂⟨H⟩/∂θᵢ for parameter i.

This method is used when GradientStrategy::Native is selected. If the backend cannot compute gradients natively, implement GradientStrategy::ParameterShiftRule instead.

Errors

Returns KetError::NativeGradientUnsuported by default. Implementations should return KetError::ExecutionFailed for backend faults.

Trait Implementations

impl Debug for dyn BatchExecution

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

Dyn Compatibility

This trait is dyn compatible.

In older versions of Rust, dyn compatibility was called "object safety".

Implementors

impl BatchExecution for CBatchExecution