# Learning Execution through Neural Code Fusion

arXiv: https://arxiv.org/abs/1906.07181 · HTML: https://mlsanity.com/paper/1906.07181/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 17 Jun 2019.

## Built on by (most cited first, 13 of 13)

- [Two Sides of the Same Coin: Heterophily and Oversmoothing in Graph Convolutional Neural Networks](https://mlsanity.com/paper/2102.06462/) (2021)
- [A Survey of Machine Learning for Computer Architecture and Systems](https://mlsanity.com/paper/2102.07952/) (2021)
- [An Evaluation of Edge TPU Accelerators for Convolutional Neural Networks](https://mlsanity.com/paper/2102.10423/) (2021)
- [Pythia: A Customizable Hardware Prefetching Framework Using Online Reinforcement Learning](https://mlsanity.com/paper/2109.12021/) (2021)
- [CodeMMLU: A Multi-Task Benchmark for Assessing Code Understanding & Reasoning Capabilities of CodeLLMs](https://mlsanity.com/paper/2410.01999/) (2024)
- [GRANITE: A Graph Neural Network Model for Basic Block Throughput Estimation](https://mlsanity.com/paper/2210.03894/) (2022)
- [Pre-Training Representations of Binary Code Using Contrastive Learning](https://mlsanity.com/paper/2210.05102/) (2022)
- [Learning Generalizable Program and Architecture Representations for Performance Modeling](https://mlsanity.com/paper/2310.16792/) (2023)
- [Program-to-Circuit: Exploiting GNNs for Program Representation and Circuit Translation](https://mlsanity.com/paper/2109.06265/) (2021)
- [Neuro-Symbolic Execution of Generic Source Code](https://mlsanity.com/paper/2304.00989/) (2023)
- [Athena: Synergizing Data Prefetching and Off-Chip Prediction via Online Reinforcement Learning](https://mlsanity.com/paper/2601.17615/) (2026)
- [NPS: A Framework for Accurate Program Sampling Using Graph Neural Network](https://mlsanity.com/paper/2304.08880/) (2023)
- [PerOS: Personalized Self-Adapting Operating Systems in the Cloud](https://mlsanity.com/paper/2404.00057/) (2024)

## Abstract (arXiv)

As the performance of computer systems stagnates due to the end of Moore's Law, there is a need for new models that can understand and optimize the execution of general purpose code. While there is a growing body of work on using Graph Neural Networks (GNNs) to learn representations of source code, these representations do not understand how code dynamically executes. In this work, we propose a new approach to use GNNs to learn fused representations of general source code and its execution. Our approach defines a multi-task GNN over low-level representations of source code and program state (i.e., assembly code and dynamic memory states), converting complex source code constructs and complex data structures into a simpler, more uniform format. We show that this leads to improved performance over similar methods that do not use execution and it opens the door to applying GNN models to new tasks that would not be feasible from static code alone. As an illustration of this, we apply the new model to challenging dynamic tasks (branch prediction and prefetching) from the SPEC CPU benchmark suite, outperforming the state-of-the-art by 26% and 45% respectively. Moreover, we use the learned fused graph embeddings to demonstrate transfer learning with high performance on an indirectly related task (algorithm classification).