.svg)
Qmod: Expressive High-Level Quantum Programming
We are excited to share that our new paper “Qmod: Expressive High-Level Quantum Programming” is now available on arXiv.
Quantum computing hardware has been advancing at breakneck speed, with major companies announcing breakthroughs in the past few months. However, there is a broad recognition that the lack of high-level programming abstractions remains a bottleneck in developing new applications. Although there are several high-level quantum programming languages out there, many still lack support for classical-like expression syntax and native constructs for basic quantum algorithmic idioms. Popular frameworks, such as Qiskit and PennyLane, essentially serve as libraries for constructing gate-level quantum circuits, making it difficult for developers to design, debug, and optimize quantum algorithms efficiently.
Qmod, developed by Classiq Technologies, is a high-level quantum programming language that enables you to capture your algorithmic intent without getting bogged down in low-level implementation details. By reintroducing conventional high-level programming concepts adapted to quantum computing semantics, Qmod significantly reduces the complexity of quantum programming. Implementation details and optimization are left to a powerful synthesis engine.
Why High-Level Abstraction Is Critical for Quantum Programming
The state of quantum programming today is reminiscent of classical computing before high-level languages like C, Java, and Python. Just as these languages revolutionized classical programming by making it more accessible and productive, Classiq (and other players in the industry) are aiming to do the same for quantum computing.
Today, most quantum programming frameworks force developers to work at the gate level, manually describing circuits and managing resources. This makes it difficult to express complex algorithms and challenging to scale applications. A journey to more evolved quantum programming languages would require the languages to incorporate concepts such as functions, variables, types, expressions, and control-flow statements. By retaining these familiar conventions and terminology, we can ease adoption and boost productivity in quantum programming.
A quantum programming language should be designed so that operations are generally reversible — with measurement as an exception — and ensure that quantum states cannot be cloned. Then, there is the issue of recycling quantum storage, which requires explicit and often costly un-computation of temporary values to disentangle auxiliary qubits and return them to their initial zero state. The scarcity of resources such as qubits, entanglement, and coherence time dramatically impacts the space of possible implementations.
To meet this challenge, Classiq Technologies created the Qmod language. Qmod captures functionality in a way that allows developers to easily communicate their intent—without worrying about implementation decisions and details. The language allows you to express quantum algorithms, and algorithmic building-blocks, without hard-coding the parameters of specific application contexts. Moreover, the compiler can adapt and optimize implementations for different hardware and simulation platforms, taking into account their constraints and restrictions.
Key Features of Qmod
Quantum functions are the basic encapsulation unit in Qmod. Just like in classical programming, decomposing algorithms into functions, or subroutines, is a key to abstraction and reuse. Quantum functions take parameters of different types, in any of the three categories – quantum types, classical types, and function types. Qmod also supports local quantum variables, quantum conditionals, and assignment statements, similar to classical programming languages. Another crucial aspect is quantum resource management. Qmod is unique in supporting functions and operations that allocate, prepare, and output quantum objects. When these operations are used in the context of the built-in conjugation construct, they must adhere to uncomputation restrictions, ensuring that qubits are implicitly uncomputed and released.
Qmod Incorporates Different Modes of Quantum Arithmetic Expressions
In quantum algorithms, qubits often encode algorithmic-level numeric data digitally in the computational basis. Arithmetic operations on digitally encoded numbers can be implemented using logic analogous to classical arithmetic, albeit reversible. However, in some situations, it is preferable to compute arithmetic operations in other ways, for example, in the Fourier basis. Quantum states can also encode information non-digitally—for example, in the amplitudes or phases of computational-basis states. Transitions between these modes of encodings are a recurring pattern in quantum algorithms.
Qmod supports native quantum expressions, combining quantum and classical variables using conventional arithmetic, relational, logic, and bitwise operators. In Qmod, quantum expressions can be evaluated in three different modes - digitally, in the phase, and in the amplitude. Amplitude and phase evaluation of expressions is a common idiom in quantum algorithms. Hence, language support for these idioms dramatically improves the conciseness and readability of the code in these cases.
Digital arithmetic works similarly to classical integer and fixed-point arithmetic, in that numbers are stored in binary format and manipulated accordingly. Qmod uses sophisticated quantum numeric type inference to optimize bit allocation in bounded precision contexts. In addition, it automatically inserts and optimizes uncomputation flow for intermediate results. Digital arithmetic is often used in quantum oracles and as a building block for general mathematical computation.
Phase arithmetic leverages the fact that quadratic expressions can be synthesized into efficient gate-level implementations. By manipulating data encoded in phases, resource requirements are significantly reduced, making computations more space-efficient. Phase arithmetic can be used, for example, to capture cost functions in optimization problems.
Amplitude arithmetic encodes the result of an expression in the amplitudes of the respective computational-basis states. This approach is based on block-encoding techniques, enabling the automatic synthesis of gate-level implementations. Algorithms for matrix inversion and Monte Carlo integration are some of the applications where amplitude-encoded expression evaluation can greatly simplify the code.
Real-World Applications of Qmod
The following two practical application examples are explained in detail in the paper, including the full code listing. For more details on the implementation, feel free to download the paper from arXiv: Qmod: Expressive High-Level Quantum Modeling.
One practical application of digital arithmetic in Qmod is approximating mathematical functions using piecewise polynomial decomposition—a task that can be implemented with straightforward code similar to classical high-level programming.
Another application example is solving combinatorial optimization problems. Qmod simplifies the formulation of optimization tasks like the integer knapsack problem with the Quantum Approximate Optimization Algorithm (QAOA). Here Qmod quantum expressions can directly represent problem objectives and constraints.
Start Building Better Quantum Programs Today
Qmod is not merely a programming language for quantum computing, it is a framework for articulating complex quantum problems, enabling more effective exploration and solution development.
The Classiq platform is a comprehensive compiler and development environment for Qmod. Using Qmod, challenging implementation tasks, such as representing intermediate and final results compactly, uncomputing intermediate results efficiently, and mapping to gate-level descriptions, are taken care of by the compiler. These capabilities address a significant bottleneck in quantum application development.
🚀 Want to learn more? Read the Qmod paper, check out Classiq’s documentation, and visit our GitHub repository.
We are excited to share that our new paper “Qmod: Expressive High-Level Quantum Programming” is now available on arXiv.
Quantum computing hardware has been advancing at breakneck speed, with major companies announcing breakthroughs in the past few months. However, there is a broad recognition that the lack of high-level programming abstractions remains a bottleneck in developing new applications. Although there are several high-level quantum programming languages out there, many still lack support for classical-like expression syntax and native constructs for basic quantum algorithmic idioms. Popular frameworks, such as Qiskit and PennyLane, essentially serve as libraries for constructing gate-level quantum circuits, making it difficult for developers to design, debug, and optimize quantum algorithms efficiently.
Qmod, developed by Classiq Technologies, is a high-level quantum programming language that enables you to capture your algorithmic intent without getting bogged down in low-level implementation details. By reintroducing conventional high-level programming concepts adapted to quantum computing semantics, Qmod significantly reduces the complexity of quantum programming. Implementation details and optimization are left to a powerful synthesis engine.
Why High-Level Abstraction Is Critical for Quantum Programming
The state of quantum programming today is reminiscent of classical computing before high-level languages like C, Java, and Python. Just as these languages revolutionized classical programming by making it more accessible and productive, Classiq (and other players in the industry) are aiming to do the same for quantum computing.
Today, most quantum programming frameworks force developers to work at the gate level, manually describing circuits and managing resources. This makes it difficult to express complex algorithms and challenging to scale applications. A journey to more evolved quantum programming languages would require the languages to incorporate concepts such as functions, variables, types, expressions, and control-flow statements. By retaining these familiar conventions and terminology, we can ease adoption and boost productivity in quantum programming.
A quantum programming language should be designed so that operations are generally reversible — with measurement as an exception — and ensure that quantum states cannot be cloned. Then, there is the issue of recycling quantum storage, which requires explicit and often costly un-computation of temporary values to disentangle auxiliary qubits and return them to their initial zero state. The scarcity of resources such as qubits, entanglement, and coherence time dramatically impacts the space of possible implementations.
To meet this challenge, Classiq Technologies created the Qmod language. Qmod captures functionality in a way that allows developers to easily communicate their intent—without worrying about implementation decisions and details. The language allows you to express quantum algorithms, and algorithmic building-blocks, without hard-coding the parameters of specific application contexts. Moreover, the compiler can adapt and optimize implementations for different hardware and simulation platforms, taking into account their constraints and restrictions.
Key Features of Qmod
Quantum functions are the basic encapsulation unit in Qmod. Just like in classical programming, decomposing algorithms into functions, or subroutines, is a key to abstraction and reuse. Quantum functions take parameters of different types, in any of the three categories – quantum types, classical types, and function types. Qmod also supports local quantum variables, quantum conditionals, and assignment statements, similar to classical programming languages. Another crucial aspect is quantum resource management. Qmod is unique in supporting functions and operations that allocate, prepare, and output quantum objects. When these operations are used in the context of the built-in conjugation construct, they must adhere to uncomputation restrictions, ensuring that qubits are implicitly uncomputed and released.
Qmod Incorporates Different Modes of Quantum Arithmetic Expressions
In quantum algorithms, qubits often encode algorithmic-level numeric data digitally in the computational basis. Arithmetic operations on digitally encoded numbers can be implemented using logic analogous to classical arithmetic, albeit reversible. However, in some situations, it is preferable to compute arithmetic operations in other ways, for example, in the Fourier basis. Quantum states can also encode information non-digitally—for example, in the amplitudes or phases of computational-basis states. Transitions between these modes of encodings are a recurring pattern in quantum algorithms.
Qmod supports native quantum expressions, combining quantum and classical variables using conventional arithmetic, relational, logic, and bitwise operators. In Qmod, quantum expressions can be evaluated in three different modes - digitally, in the phase, and in the amplitude. Amplitude and phase evaluation of expressions is a common idiom in quantum algorithms. Hence, language support for these idioms dramatically improves the conciseness and readability of the code in these cases.
Digital arithmetic works similarly to classical integer and fixed-point arithmetic, in that numbers are stored in binary format and manipulated accordingly. Qmod uses sophisticated quantum numeric type inference to optimize bit allocation in bounded precision contexts. In addition, it automatically inserts and optimizes uncomputation flow for intermediate results. Digital arithmetic is often used in quantum oracles and as a building block for general mathematical computation.
Phase arithmetic leverages the fact that quadratic expressions can be synthesized into efficient gate-level implementations. By manipulating data encoded in phases, resource requirements are significantly reduced, making computations more space-efficient. Phase arithmetic can be used, for example, to capture cost functions in optimization problems.
Amplitude arithmetic encodes the result of an expression in the amplitudes of the respective computational-basis states. This approach is based on block-encoding techniques, enabling the automatic synthesis of gate-level implementations. Algorithms for matrix inversion and Monte Carlo integration are some of the applications where amplitude-encoded expression evaluation can greatly simplify the code.
Real-World Applications of Qmod
The following two practical application examples are explained in detail in the paper, including the full code listing. For more details on the implementation, feel free to download the paper from arXiv: Qmod: Expressive High-Level Quantum Modeling.
One practical application of digital arithmetic in Qmod is approximating mathematical functions using piecewise polynomial decomposition—a task that can be implemented with straightforward code similar to classical high-level programming.
Another application example is solving combinatorial optimization problems. Qmod simplifies the formulation of optimization tasks like the integer knapsack problem with the Quantum Approximate Optimization Algorithm (QAOA). Here Qmod quantum expressions can directly represent problem objectives and constraints.
Start Building Better Quantum Programs Today
Qmod is not merely a programming language for quantum computing, it is a framework for articulating complex quantum problems, enabling more effective exploration and solution development.
The Classiq platform is a comprehensive compiler and development environment for Qmod. Using Qmod, challenging implementation tasks, such as representing intermediate and final results compactly, uncomputing intermediate results efficiently, and mapping to gate-level descriptions, are taken care of by the compiler. These capabilities address a significant bottleneck in quantum application development.
🚀 Want to learn more? Read the Qmod paper, check out Classiq’s documentation, and visit our GitHub repository.
About "The Qubit Guy's Podcast"
Hosted by The Qubit Guy (Yuval Boger, our Chief Marketing Officer), the podcast hosts thought leaders in quantum computing to discuss business and technical questions that impact the quantum computing ecosystem. Our guests provide interesting insights about quantum computer software and algorithm, quantum computer hardware, key applications for quantum computing, market studies of the quantum industry and more.
If you would like to suggest a guest for the podcast, please contact us.
