1. MARCO - Modelica Advanced Research COmpiler - Michele Scuttari
2. Translation Validation for LLVM's RISC-V Backend - Mitch Briles
3. Leveraging MLIR to Compile a Basis-Oriented Quantum Programming Language - Austin J. Adams

MARCO - Modelica Advanced Research COmpiler - Michele Scuttari
Modelica is a high-level, equation-based language used for modeling complex physical systems. MARCO (Modelica Advanced Research Compiler) is a novel compiler that leverages MLIR to bring modern compiler infrastructure to the Modelica ecosystem. This talk introduces MARCO’s architecture and its MLIR-based lowering strategy for Modelica's Differential-Algebraic Equation (DAE) systems. The talk is aimed at developers interested in domain-specific language compilation, and attendees will take away insights on integrating MLIR for descriptive, non-traditional, languages like Modelica. The project has been first and only presented at the Modelica conference in 2023, but it has never been proposed to a compiler-oriented audience.

Translation Validation for LLVM's RISC-V Backend - Mitch Briles
With algorithms such as instruction selection, instruction folding, and register allocation, LLVM's backends have the job of lowering IR to assembly or object code while being mindful of the semantics of each language. Starting with AArch64, we're leveraging Alive2 to validate these target-dependent optimizations and translations. After using our tool to find 44 miscompiles in the AArch64 backend, the natural progression is to branch out to other architectures. Our next focus is a much smaller ISA: RISC-V. The new tool, RISCV-TV, is early in development, but has already detected 2 miscompiles! Bugs can be found in existing tests, but these tools are most effective when paired with a fuzzer. We anticipate more results by the time of the meeting.

Leveraging MLIR to Compile a Basis-Oriented Quantum Programming Language - Austin J. Adams
Quantum computing has leaped from the theoretical realm into a competitive commercial race, but programming quantum computers remains challenging. Quantum programming languages today require programmers to master both physics notation and quantum gate engineering. The Qwerty quantum programming language was recently proposed (arXiv:2404.12603) as a higher-level abstraction providing primitives rooted in quantum bases rather than low-level quantum circuitry. The semantic gap between Qwerty's novel basis-oriented constructs and traditional quantum circuits renders existing quantum compilers unsuitable for compiling Qwerty code. This talk describes how MLIR is used in our Qwerty compiler to compile Qwerty code embedded in Python into quantum assembly or LLVM IR. The full architecture of our compiler is described, including custom MLIR dialects, interfaces, analyses, and passes. A brief comparison between all known quantum MLIR dialects will also be presented.