CodeJeNN: A simple C++ neural network generator for physics applications

Jay Arcities; Pavel Popov; Eric J Ching; Kamal Viswanath; Ryan F Johnson

physics.comp-phJul 2, 2026

CodeJeNN: A simple C++ neural network generator for physics applications

Authors: Jay Arcities, Pavel Popov, Eric J Ching, Kamal Viswanath, Ryan F Johnson

Organizations: San Diego State University, Aerospace Engineering Department, San Diego, CA 92182 · U.S. Naval Research Laboratory, Laboratories of Computational Physics and Fluid Dynamics, Washington, DC 20375

Abstract

Machine learning has shown speedups for numerical methods in physics applications, but integrating Python-based libraries into high-performance C++ solvers creates performance bottlenecks. We present CodeJeNN, which bridges this gap by auto-generating self-contained C++ code from trained Keras models for inference. This eliminates external dependencies through minimal inlined functions, allowing seamless integration into existing frameworks. We describe the Keras-to-C++ workflow, supported architectures, and limitations. CodeJeNN is demonstrated through inference benchmarks against Keras in eager and JIT modes and a CFD test case modeling viscosity in a hydrogen-air mixing layer, showing speedups without sacrificing accuracy.

Explore similar work

Beyond the Need for Speed: Energy-Aware Code Generation via Simulation-Guided Reinforcement Learning
Jul 6, 2026Saurabhsingh Rajput, Tushar SharmaEnergy-Efficient Code Generation
KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators
Jun 1, 2026Taras Sereda, Burak Bartan, Ankita Nayak +3Graphics Processing Unit Kernels Accelerator
AgentCompile: An LLM-Guided Compiler for Direct CUDA Inference
Jun 4, 2026Xuanzhe Li, Ziyan Weng, Zhiyu Zhu +1Cuda Pytorch

Jul 6, 2026 · cs.LGJ/K move · Enter open · S save

Beyond the Need for Speed: Energy-Aware Code Generation via Simulation-Guided Reinforcement Learning

Saurabhsingh Rajput, Tushar Sharma

Code models strictly prioritize functional correctness, leaving software energy efficiency as an unoptimized byproduct. Training models to generate energy-efficient code requires reproducible feedback at scale, which physical hardware measurement cannot reliably provide due to variance. In this paper, we replace hardware profiling with a deterministic architectural simulation harness to build Green Tea, a corpus of

3.5

million evaluations across

1{,}474

C++ problems. We train an energy-aware code model via supervised fine-tuning on energy-contrastive pairs, followed by closed-loop reinforcement learning (GRPO) using simulation-in-the-loop feedback. To rigorously evaluate deployment readiness, we introduce the Correctness-Adjusted Reduction in Energy Total (CARET), a metric that explicitly penalizes code that sacrifices functionality for efficiency. On

143

held-out problems, our simulation-in-the-loop pipeline achieves

12.63\%

CARET, nearly tripling the gain of fine-tuning alone, and successfully beats the energy efficiency of human-expert references on

58.4\%

of its valid outputs. Furthermore, our analysis exposes the IPC trap: standard throughput proxies like Instructions-Per-Cycle (IPC) actively misrank true energy efficiency on

67.8\%

of problems, proving the absolute necessity of direct energy simulation. By releasing our dataset and infrastructure, we bypass the

263{,}000

CPU-hours required for reproduction, structurally empowering the community to deploy inherently energy-efficient code generation models.

Energy-Efficient Code Generation Profiling

Open paper Ask about it PDF

Abstract

Explore similar work

Beyond the Need for Speed: Energy-Aware Code Generation via Simulation-Guided Reinforcement Learning

KForge: LLM-Driven Cross-Platform Kernel Generation for AI Accelerators

AgentCompile: An LLM-Guided Compiler for Direct CUDA Inference