What's New

Changes from 2.1.5 to 3.1

Efficient functions for checking equivalence up to global and relative phase of operators and states. The phase factors are also computed by most of these functions. See the User's Guide for details. The algorithms that these functions are based on are described in George Viamontes' thesis. Example scripts may be found in examples/phase_checks/
Support for Dirac-style notation. See the User's Guide for examples and details. Example scripts may be found in examples/states/
Support for string data types, including variable storage and string manipulation, has been added. The User's Guide has examples of this feature. This feature also makes the cu_gate function much easier to use.
New operators including +=, -=, *=, and /=
Bit manipulation operators including << and >>. Also functions to get and set individual bits of an integer variable have been added.
Binary integer expressions are now supported using the bn syntax.
The abs function has been added, which implements the complex modulus operation.
Alternate syntax for the kron function in the form of the (X) operator, which makes writing Dirac-style expressions more natural.
Alternate syntax for the hadamard function. H is now syntactic short-hand for it.
Changed the matrix produced by sigma_z. It differs by a mutiple of -1 from the old matrix produced. This is only a global phase change, but several requests indicated that this version would be preferred.

A bug which degraded runtime and memory performance when using loops and function calls within tight loops has been fixed.
In some cases when multiplying a call to cu_gate with a state vector, the state vector was modified incorrectly. This bug has been fixed.
Fixed the cnot and toffoli functions, which previously would incorrectly modify state vectors in certain cases.

Fixed a bug in which variations of Toffoli gates (i.e. controlled-NOT gates with two controls) would not modify state vectors properly.
Improved numerical precision.

Fixed a bug in which lazily created cu_gate operators were all treated as controlled-NOT gates even if the action specified in the cu_gate call was not X.

Improved performance on a few benchmarks, particularly grover_density.qpro, by changing the DD hashing function of terminal values.
Fixed a bug which caused some imaginary values of complex numbers within a matrix or vector to be off by -i or take on slightly incorrect values. This fix in particular affected calls to the rz function.

Fixed a bug which caused imaginary values of complex numbers within a matrix or vector to be set to 0 when the real value was 0.

Fixed a bug which periodically caused the parser to crash when certain calls to the state vector normalization function were made.
Fixed a bug which caused incorrect values to be generated for state vector normalization.

Support for user-defined functions has been added. See Section 4 of the User's Guide for details.
The run command has been added which executes the script code contained in a separate file. This command is ideal for declaring variables which may be re-used in multiple projects. See Section 4 of the User's Guide for details.

A swap function has been added which efficiently swaps two qubits in a state vector. This function is much faster than swapping the qubits explicitly using CNOT and Hadamard gates.
A dump_dot function has been added which outputs a file that can be converted into the graphical representation of a QuIDD. Making use of this function requires the freely available Graphviz package which can be obtained at http://www.graphviz.org/
A quidd_info function has been added which outputs information about an operator or density matrix. See Section 4 of the User's Guide for more details.

Column entries in a matrix may be separated by commas now in addition to whitespace.

Examples have been added demonstrating user-defined functions and the run command.