ParaCodex: A Profiling-Guided Autonomous Coding Agent for Reliable Parallel Code Generation and Translation

2026-01-07

arXiv.org

AI Summary

This work addresses the challenges of data movement and performance tuning in OpenMP GPU offloading by proposing a Codex-based autonomous coding agent. The agent employs a domain-knowledge-guided workflow that integrates hotspot analysis, explicit data layout planning, correctness gating, and performance-profile-driven iterative optimization to automatically translate sequential CPU code into efficient and reliable OpenMP GPU offload code. Evaluated on 31 kernels, the approach successfully generates runnable code for all cases, with 25 outperforming reference implementations. It achieves geometric mean speedups of 3× and 5× on the HeCBench and Rodinia benchmarks, respectively, and demonstrates high compilation and validation success rates in CUDA-to-OpenMP migration tasks.

Technology Category

• Search and Optimization: Sampling/Simulation-based Search
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization

Abstract

Parallel programming is central to HPC and AI, but producing code that is correct and fast remains challenging, especially for OpenMP GPU offload, where data movement and tuning dominate. Autonomous coding agents can compile, test, and profile on target hardware, but outputs are brittle without domain scaffolding. We present ParaCodex, an HPC-engineer workflow that turns a Codex-based agent into an autonomous OpenMP GPU offload system using staged hotspot analysis, explicit data planning, correctness gating, and profiling-guided refinement. We evaluate translation from serial CPU kernels to OpenMP GPU offload kernels on HeCBench, Rodinia, and NAS. After excluding five kernels, ParaCodex succeeded on all 31 valid kernels. The generated kernels improved GPU time over reference OpenMP implementations in 25/31 cases, achieving geometric-mean speedups of 3x on HeCBench and 5x on Rodinia, and outperforming a zero-shot Codex baseline on all suites. We also evaluate CUDA to OpenMP offload translation on ParEval, where ParaCodex maintains high compilation and validation rates in code-only and end-to-end settings.

Problem

• parallel programming
• OpenMP GPU offload
• code correctness
• performance optimization
• autonomous coding

Innovation

• Profiling-Guided Code Generation
• Autonomous Coding Agent
• OpenMP GPU Offload
• Correctness Gating
• Data Movement Planning