Compilation passes

Compilation passes are essential steps in the process of converting a high-level quantum algorithm, i.e. quantum circuits, into a hardware-specific executable format. This process, known as quantum compilation, involves several stages to ensure that the quantum circuit can be executed efficiently on a given quantum hardware.

Compilation passes include various optimization techniques and transformations applied to the quantum circuit. These passes can involve qubit routing, initial mapping, gate decomposition, or error correction. These optimization steps are essential to the execution of quantum algorithms. Often times, the design of quantum algorithms does not take into account the constraints or limitations imposed by the target hardware, such as qubit coupling map or native gate set.

These passes are therefore needed to ensure that an initial circuit is converted to a version that adheres to the requirements of the hardware. They can easily be applied using the following methods on the circuit object:

• decompose
• export
• map
• merge
• route
• validate

Types of passes

Given the methods stated above, the following types of passes are available:

• Decomposer
• Exporter
• Mapper
• Merger
• Router
• Validator

Integrated passes

The reader and writer passes are integrated in particular functionalities of the circuit. They are not applied in the same way as the passes mentioned above, i.e., by passing them as an argument when calling one of the aforementiond methods on the circuit. Instead, the reader and writer are executed when one parses a cQASM string or writes out the circuit to a cQASM string, respectively. The reader is invoked when using the Circuit.from_string method, and the writer is invoked when converting the circuit to a string with str or printing it out with print.