tqec.simulation.simulation.start_simulation_using_sinter#

start_simulation_using_sinter(block_graph: ~tqec.computation.block_graph.BlockGraph, ks: ~typing.Sequence[int], ps: ~typing.Sequence[float], noise_model_factory: ~typing.Callable[[float], ~tqec.noise_models.noise_model.NoiseModel], manhattan_radius: int, block_builder: ~tqec.compile.specs.base.BlockBuilder = functools.partial(<function default_compiled_block_builder>, plaquette_builder=<function make_css_surface_code_plaquette>), substitution_builder: ~tqec.compile.specs.base.SubstitutionBuilder = functools.partial(<function default_substitution_builder>, plaquette_builder=<function make_css_surface_code_plaquette>), observables: list[~tqec.computation.abstract_observable.AbstractObservable] | None = None, num_workers: int = 4, progress_callback: ~typing.Callable[[~sinter._collection._collection.Progress], None] | None = None, max_shots: int | None = None, max_errors: int | None = None, decoders: ~typing.Iterable[str] = ('pymatching',), print_progress: bool = False, custom_decoders: dict[str, ~sinter._decoding._decoding_decoder_class.Decoder | ~sinter._decoding._sampler.Sampler] | None = None) → Iterator[list[TaskStats]][source]#

Helper to run stim simulations using sinter.

This function is the preferred entry-point to run sinter computations using stim from circuits generated by tqec.

It remove the need to generate the stim.Circuit instances in parallel (due to the relative inefficiency of detector annotation at the moment) by importing and calling generate_sinter_tasks(). It also forwards several parameters to sinter.collect(), but without showing in its signature arguments that we do not envision using.

Parameters:

  • block_graph – a representation of the QEC computation to simulate.

  • ks – values of the scaling parameter k to use in order to generate the circuits.

  • ps – values of the noise parameter p to use to instantiate a noise model using the provided noise_model_factory.

  • noise_model_factory – a callable that is used to instantiate a noise model from each of the noise parameters in ps.

  • manhattan_radius – radius used to automatically compute detectors. The best value to set this argument to is the minimum integer such that flows generated from any given reset/measurement, from any plaquette at any spatial/temporal place in the QEC computation, do not propagate outside of the qubits used by plaquettes spatially located at maximum manhattan_radius plaquettes from the plaquette the reset/measurement belongs to (w.r.t. the Manhattan distance). Default to 2, which is sufficient for regular surface code. If negative, detectors are not computed automatically and are not added to the generated circuits.

  • block_builder – A callable that specifies how to build the CompiledBlock from the specified CubeSpecs. Defaults to the block builder for the css type surface code.

  • substitution_builder – A callable that specifies how to build the substitution plaquettes from the specified PipeSpec. Defaults to the substitution builder for the css type surface code.

  • observables – a list of observables to generate statistics for. If None, all the observables of the provided computation are used.

  • num_workers – The number of worker processes to use.

  • progress_callback – Defaults to None (unused). If specified, then each time new sample statistics are acquired from a worker this method will be invoked with the new sinter.TaskStats.

  • max_shots – Defaults to None (unused). Stops the sampling process after this many samples have been taken from the circuit.

  • max_errors – Defaults to None (unused). Stops the sampling process after this many errors have been seen in samples taken from the circuit. The actual number sampled errors may be larger due to batching.

  • decoders – Defaults to None (specified by each Task). The names of the decoders to use on each Task. It must either be the case that each Task specifies a decoder and this is set to None, or this is an iterable and each Task has its decoder set to None.

  • print_progress – When True, progress is printed to stderr while collection runs.

  • custom_decoders – Named child classes of sinter.decoder, that can be used if requested by name by a task or by the decoders list. If not specified, only decoders with support built into sinter, such as ‘pymatching’ and ‘fusion_blossom’, can be used.

Yields:

one simulation result (of type list[sinter.TaskStats]) per provided observable in observables.