Abstract:Code translation transforms programs from one programming language (PL) to another. Several rule-based transpilers have been designed to automate code translation between different pairs of PLs. However, the rules can become obsolete as the PLs evolve and cannot generalize to other PLs. Recent studies have explored the automation of code translation using Large Language Models (LLMs). One key observation is that such techniques may work well for crafted benchmarks but fail to generalize to the scale and complexity of real-world projects with dependencies, custom types, PL-specific features, etc. We propose AlphaTrans, a neuro-symbolic approach to automate repository-level code translation. AlphaTrans translates both source and test code, and employs multiple levels of validation to ensure the translation preserves the functionality of the source program. To break down the problem for LLMs, AlphaTrans leverages program analysis to decompose the program into fragments and translates them in the reverse call order. We leveraged AlphaTrans to translate ten real-world open-source projects consisting of <836, 8575, 2719> classes, methods, and tests. AlphaTrans translated the entire repository of these projects consisting of 6899 source code fragments. 99.1% of the translated code fragments are syntactically correct, and AlphaTrans validates the translations' runtime behavior and functional correctness for 25.8%. On average, the integrated translation and validation take 36 hours to translate a project, showing its scalability in practice. For the syntactically or semantically incorrect translations, AlphaTrans generates a report including existing translation, stack trace, test errors, or assertion failures. We provided these artifacts to two developers to fix the translation bugs in four projects. They were able to fix the issues in 20.1 hours on average and achieve all passing tests.

Trustworthy Translation for Eliminating High-Order Operation of a Synchronous Dataflow Language

Key Translations of the Trustworthy Compiler L2C and Its Design and Implementation

Credible sorting for synchronous data-flow programs

A Formally Verified Sequentializer for Lustre-Like Concurrent Synchronous Data-Flow Programs

A Formally Verified Transformation to Unify Multiple Nested Clocks for a Lustre-like Language

Verified Lustre Normalization with Node Subsampling

Design and FPGA Verification of a Novel Reliable Real-Time Data Transfer System

Verified Code Transpilation with LLMs

The synchronous data flow programming language LUSTRE

Towards Translating Real-World Code with LLMs: A Study of Translating to Rust

WaveCert: Translation Validation for Asynchronous Dataflow Programs via Dynamic Fractional Permissions

VERT: Verified Equivalent Rust Transpilation with Large Language Models as Few-Shot Learners

Enhancing Translation Validation of Compiler Transformations with Large Language Models

LASSI: An LLM-based Automated Self-Correcting Pipeline for Translating Parallel Scientific Codes

Sound Transpilation from Binary to Machine-Independent Code

Towards certified separate compilation for concurrent programs

Towards Proved Formal Specification and Verification of STL Operators as Synchronous Observers

A Systematic Translation Validation Framework for MLIR-Based Compilers

Scalable, Validated Code Translation of Entire Projects using Large Language Models

Repository-Level Compositional Code Translation and Validation

LightningSim: Fast and Accurate Trace-Based Simulation for High-Level Synthesis