Work

Compiling Linear Algebra Workloads from C to Quantum Circuits via Multi Level Intermediate Representation

Year: 2026

Type: conference-paper

Abstract: Linear algebra kernels are central to many quantum algorithms, yet existing quantum compilers primarily target scalar arithmetic, limiting their expressiveness and applicability to such workloads. In this work, we extend the QuantumC compilation framework to support the compilation of C programs containing linear algebra operations into quantum circuits. Our approach introduces pattern-based recognition of common vector and matrix computations, including vector addition, dot product, and matrix–matrix multiplication, and integrates a dedicated intermediate representation into the existing QuantumC pipeline. This design allows selecting between two quantum arithmetic backends—ripple-carry and quantum Fourier transform—enabling trade-offs in quantum resource usage. Additionally, we introduce a compile-time sparsity optimization pass that statically analyzes vector and matrix structures. For matrices with \(90\%\) sparsity in matrix multiplication, this optimization reduces qubit count by 91–\(97\%\), T-depth by 85–\(99\%\), and T-count by \(99\%\), depending on the selected arithmetic backend.

Source: ACM International Conference on Computing Frontiers: Workshops and Special Sessions (CF Companion)

DOI: 10.1145/3801488.3806071

Authors: Salvatore Pio Conte, Simone Cosenza, Giacomo Lancellotti, Giovanni Agosta

Institution: Politecnico di Milano

Language: English

Topic: Quantum Computing Algorithms and Architecture

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: hybrid

OpenAlex ID: W7167940001

FWCI:

Cites: 19

Cited by:

Related to: 10