pygsti.layouts.matrixlayout
¶
Defines the MatrixCOPALayout class.
Module Contents¶
Classes¶
The atom ("atomic unit") for dividing up the element dimension in a 

A circuit outcome probability array (COPA) layout for circuit simulation by process matrix multiplication. 
 class pygsti.layouts.matrixlayout._MatrixCOPALayoutAtom(unique_complete_circuits, unique_nospam_circuits, circuits_by_unique_nospam_circuits, ds_circuits, group, helpful_scratch, model, dataset)¶
Bases:
pygsti.layouts.distlayout._DistributableAtom
The atom (“atomic unit”) for dividing up the element dimension in a
MatrixCOPALayout
. Parameters
unique_complete_circuits (list) – A list that contains all the “complete” circuits for the parent layout. This atom only owns a subset of these, as given by group below.
unique_nospam_circuits (list) – A list that contains the unique circuits within unique_complete_circuits once their state preparations and measurements are removed. A subset of these circuits (see group below) are what fundamentally define the circuit outcomes that this atom includes: it includes all the circuit outcomes of those circuits.
circuits_by_unique_nospam_circuits (dict) – A dictionary with keys equal to the elements of unique_nospam_circuits and values that are lists of indices into unique_complete_circuits. Thus, this dictionary maps each distinct circuitwithoutSPAM circuit to the list of complete circuits within unique_complete_circuits that correspond to it.
ds_circuits (list) – A list of circuits parallel to unique_complete_circuits of these circuits as they should be accessed from dataset. This applies any aliases and removes implied SPAM elements relative to unique_complete_circuits.
group (set) – The set of indices into unique_nospam_circuits that define the circuit outcomes owned by this atom.
helpful_scratch (set) – A set of indices into unique_nospam_circuits that specify circuits that aren’t owned by this atom but are helpful in building up an efficient evaluation tree.
model (Model) – The model being used to construct this layout. Used for expanding instruments within the circuits.
dataset (DataSet) – The dataset, used to include only observed circuit outcomes in this atom and therefore the parent layout.
 nonscratch_cache_view(self, a, axis=None)¶
Create a view of array a restricting it to only the final results computed by this tree.
This need not be the entire array because there could be intermediate results (e.g. “scratch space”) that are excluded.
 Parameters
a (ndarray) – An array of results computed using this EvalTree, such that the axisth dimension equals the full length of the tree. The other dimensions of a are unrestricted.
axis (int, optional) – Specified the axis along which the selection of the final elements is performed. If None, than this selection if performed on flattened a.
 Returns
ndarray – Of the same shape as a, except for along the specified axis, whose dimension has been reduced to filter out the intermediate (nonfinal) results.
 property cache_size(self)¶
The cache size of this atom.
 class pygsti.layouts.matrixlayout.MatrixCOPALayout(circuits, model, dataset=None, num_sub_trees=None, num_tree_processors=1, num_param_dimension_processors=(), param_dimensions=(), param_dimension_blk_sizes=(), resource_alloc=None, verbosity=0)¶
Bases:
pygsti.layouts.distlayout.DistributableCOPALayout
A circuit outcome probability array (COPA) layout for circuit simulation by process matrix multiplication.
A distributed layout that divides a list of circuits into several “evaluation trees” that compute subsets of the circuit outcomes by multiplying together process matrices. Often these evaluation trees correspond to available processors, but it can be useful to divide computations in order to lessen the amount of intermediate memory required.
MatrixCOPALayout instances create and store the decomposition of a list of circuits into a sequence of 2term products of smaller strings. Ideally, this sequence would prescribe the way to obtain the entire list of circuits, starting with just the single gates, using the fewest number of multiplications, but this optimality is not guaranteed.
 Parameters
circuits (list) – A list of:class:Circuit objects representing the circuits this layout will include.
model (Model) – The model that will be used to compute circuit outcome probabilities using this layout. This model is used to complete and expand the circuits in circuits.
dataset (DataSet, optional) – If not None, restrict the circuit outcomes stored by this layout to only the outcomes observed in this data set.
num_sub_trees (int, optional) – The number of groups (“subtrees”) to divide the circuits into. This is the number of atoms for this layout.
num_tree_processors (int, optional) – The number of atomprocessors, i.e. groups of processors that process subtrees.
num_param_dimension_processors (tuple, optional) – A 1 or 2tuple of integers specifying how many parameterblock processors are used when dividing the physical processors into a grid. The first and second elements correspond to counts for the first and second parameter dimensions, respecively.
param_dimensions (tuple, optional) – The number of parameters along each parameter dimension. Can be an empty, 1, or 2tuple of integers which dictates how many parameter dimensions this layout supports.
param_dimension_blk_sizes (tuple, optional) – The parameter block sizes along each present parameter dimension, so this should be the same shape as param_dimensions. A block size of None means that there should be no division into blocks, and that each block processor computes all of its parameter indices at once.
resource_alloc (ResourceAllocation, optional) – The resources available for computing circuit outcome probabilities.
verbosity (int or VerbosityPrinter) – Determines how much output to send to stdout. 0 means no output, higher integers mean more output.