Skip to content

Latest commit

 

History

History
58 lines (47 loc) · 4.11 KB

cpp_solver_interface.md

File metadata and controls

58 lines (47 loc) · 4.11 KB
Raw

The C++ solver interface

This file describes the low-level C++ interface. Typical usage of the package does not require following most of the details described below. The recommended usage is with Python, where the wrapper generates the input file for the solver in the correct format.

One notable use-case is when running the solver executable on a remote server (or cluster node) with an input file generated using the Python interface on a different machine. In this case, follow the steps below after generaring an inout file using the Python interface:

  • Update the paths inserted in the input file to be those relevant for target machine.
  • Run the solver from the command line. The solver executable expects a single parameter in its command line, which is the string literal input_file, followed by a space and the file name of an input file. See the installation instructions for guaranteeing the proper multithreaded BLAS/LAPACK libraries are available.
  • When the solver terminates, use the Python interface to load the output data files and plot the results.

C++ solver input files

In the input file each parameter is specified on a separate line, which contains the parameter name, a space, the equality sign '=', a space, and a single value or a comma-separated list of values (without spaces).

The solver that is based on the ITesnor library, uses 1-based indexes for the qubits, and this is reflected in the input and output files. The Python code converts the 0-based indexes that are used by convention in Python indexing when creating the input file and when loading the output files of the solver, so that the Python interface exposes 0-based indexes.

The parameters that the executable accepts have identical names and meaning as the parameters detailed for the Python interface, with a few exceptions detailed in the following.

  • b_unique_id. This parameter is not recognized by the solver.
  • unique_id, metadata. These string parameters are currently ignored by the solver, except for being allowed in the input file and saved in the log file.
  • J, J_z. These two executable parameters are passed as a single scalar value (if the Python parameter is scalar), or as a list (no matrix format is supported). If either J or J_z is a matrix in the Python, then the list consists of all entries which are nonzero in either one of the two matrices. In this case, there are two additional parameters used in order to describe the qubit pairs to which the values correspond;
  • first_bond_indices, second_bond_indices. Lists of the indices of the first and second qubit (respectively) of each entry in the lists J, J_z (that must be of identical length).

C++ solver output files

The solver executable generates files: a saved state (at the simulation end), observables at requested times, global solver data at the same times as the observables, and a log file. The format of the saved state file is not described here. The log file is textual. The observabl files have a consistent table format with each line holding one data value with tab-separated fields, as detailed in the following. We note that similarly to the input file, the output files qubit indexes are 1-based.

The three output files are:

  • The one-qubit observables file, with the following structure as defined by the columns: time, operator, index, value. The time column stores the time at which the observable is taken, operator is the Pauli operator {X,Y,Z}, index is the qubit index and value is the Pauli expectation value.

  • The two-qubit observables file has the columns: time, operator, index_1, index_2, value. The time column stores the time at which the observable is taken, operator is the Pauli operator XX, XY, etc., index_1 is the first qubit index, index_2 is the second qubit index and value is the two-qubit Pauli expectation value.

  • The global quantities file has the columns: time, quantity, value. The time column stores the time at which the quantity is taken, quantity is a string denoting the name of the calculated global quantity and value is the value.