Cobble: Compiling Block Encodings for Quantum Computational Linear Algebra
summary
Cobble compiles high-level matrix expressions into correct quantum circuits by representing matrices as block encodings, computing time and space costs including the crucial alpha subnormalization factor, and applying program-level rewrites plus the quantum singular value transformation to shrink circuits and reduce repetition. The compiler emits circuits that beat unoptimized baselines (2.6×–25.4× faster on benchmarks) and often outperform generic circuit optimizers, with a reusable artifact package available for inspection.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:06Quantum algorithms for linear algebra
- 3:14Quantum programming challenges
- 6:22Quantum linear algebra
- 9:06High-level quantum language design
- 11:32Encoding matrices on quantum computers
- 14:39Measuring and reducing program costs
- 17:49Optimizing circuits with PL techniques
- 18:32Call to action for quantum software
- 19:14Q&A
related talks
22mProgramming Abstractions for Quantum Linear Algebra
39mBuilding a Compiler for Quantum Computers
21mToward Certified Quantum Programming
18mAutomated distribution of quantum circuits via hypergraph partitioning
31mQuantum compilation using the ZX calculus
14mDependency-Aware Compilation for Surface Code Quantum Architectures
49mQuantum Garbled Circuits
15mToward A Quantum Programming Language for Higher Level Formal Verification