cotengra.presets

Preset configured optimizers.

Attributes

greedy_optimize
optimal_optimize
optimal_outer_optimize
edgesort_optimize

Classes

AutoOptimizer	An optimizer that automatically chooses between optimal and
AutoHQOptimizer	An optimizer that automatically chooses between optimal and

Functions

estimate_optimal_hardness(inputs)	Provides a very rough estimate of how long it would take to find the
get_auto_optimizer()	Cache and reuse for speed. But only construct dynamically.
auto_optimize(inputs, output, size_dict, **kwargs)
auto_optimize_tree(inputs, output, size_dict, **kwargs)	Get a contraction tree using the auto optimizer.
get_auto_hq_optimizer()	Cache and reuse for speed. But only construct dynamically.
auto_hq_optimize(inputs, output, size_dict, **kwargs)
auto_optimize_hq_tree(inputs, output, size_dict, **kwargs)	Get a contraction tree using the auto optimizer.

Module Contents

cotengra.presets.estimate_optimal_hardness(inputs)

Provides a very rough estimate of how long it would take to find the optimal contraction order for a given set of inputs. The runtime is very approximately exponential in this number:

\[T \propto \exp {n^2 * k^0.5}\]

Where \(n\) is the number of tensors and \(k\) is the average degree of the hypergraph.

class cotengra.presets.AutoOptimizer(optimal_cutoff=250, minimize='combo', cache=True, **hyperoptimizer_kwargs)

Bases: cotengra.oe.PathOptimizer

An optimizer that automatically chooses between optimal and hyper-optimization, designed for everyday use.

minimize = 'combo'

optimal_cutoff = 250

_optimize_optimal_fn

kwargs

_hyperoptimizers_by_thread

_get_optimizer_hyper_threadsafe()

search(inputs, output, size_dict, **kwargs)

__call__(inputs, output, size_dict, **kwargs)

class cotengra.presets.AutoHQOptimizer(**kwargs)

Bases: AutoOptimizer

An optimizer that automatically chooses between optimal and hyper-optimization, designed for everyday use on harder contractions or those that will be repeated many times, and thus warrant a more extensive search.

cotengra.presets.get_auto_optimizer()

Cache and reuse for speed. But only construct dynamically.

cotengra.presets.auto_optimize(inputs, output, size_dict, **kwargs)

cotengra.presets.auto_optimize_tree(inputs, output, size_dict, **kwargs)

Get a contraction tree using the auto optimizer.

cotengra.presets.get_auto_hq_optimizer()

Cache and reuse for speed. But only construct dynamically.

cotengra.presets.auto_hq_optimize(inputs, output, size_dict, **kwargs)

cotengra.presets.auto_optimize_hq_tree(inputs, output, size_dict, **kwargs)

Get a contraction tree using the auto optimizer.

cotengra.presets.greedy_optimize

cotengra.presets.optimal_optimize

cotengra.presets.optimal_outer_optimize

cotengra.presets.edgesort_optimize