Advising OpenMP Parallelization via A Graph-Based Approach with Transformers

Tal Kadosh, Nadav Schneider, Niranjan Hasabnis, Timothy Mattson, Yuval Pinter, Gal Oren

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review


There is an ever-present need for shared memory parallelization schemes to exploit the full potential of multi-core architectures. The most common parallelization API addressing this need today is OpenMP. Nevertheless, writing parallel code manually is complex and effort-intensive. Thus, many deterministic source-to-source (S2S) compilers have emerged, intending to automate the process of translating serial to parallel code. However, recent studies have shown that these compilers are impractical in many scenarios. In this work, we combine the latest advancements in the field of AI and natural language processing (NLP) with the vast amount of open-source code to address the problem of automatic parallelization. Specifically, we propose a novel approach, called OMPify, to detect and predict the OpenMP pragmas and shared-memory attributes in parallel code, given its serial version. OMPify is based on a Transformer-based model that leverages a graph-based representation of source code that exploits the inherent structure of code. We evaluated our tool by predicting the parallelization pragmas and attributes of a large corpus of (over 54,000) snippets of serial code written in C and C++ languages (Open-OMP-Plus). Our results demonstrate that OMPify outperforms existing approaches — the general-purposed and popular ChatGPT and targeted PragFormer models — in terms of F1 score and accuracy. Specifically, OMPify achieves up to 90% accuracy on commonly-used OpenMP benchmark tests such as NAS, SPEC, and PolyBench. Additionally, we performed an ablation study to assess the impact of different model components and present interesting insights derived from the study. Lastly, we also explored the potential of using data augmentation and curriculum learning techniques to improve the model’s robustness and generalization capabilities. The dataset and source code necessary for reproducing our results are available at

Original languageEnglish
Title of host publicationOpenMP
Subtitle of host publicationAdvanced Task-Based, Device and Compiler Programming - 19th International Workshop on OpenMP, IWOMP 2023, Proceedings
EditorsSimon McIntosh-Smith, Tom Deakin, Michael Klemm, Bronis R. de Supinski, Jannis Klinkenberg
PublisherSpringer Science and Business Media Deutschland GmbH
Number of pages15
ISBN (Print)9783031407437
StatePublished - 1 Jan 2023
EventProceedings of the 19th International Workshop on OpenMP, IWOMP 2023 - Bristol, United Kingdom
Duration: 13 Sep 202315 Sep 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14114 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


ConferenceProceedings of the 19th International Workshop on OpenMP, IWOMP 2023
Country/TerritoryUnited Kingdom


  • Code Completion
  • Code Representations
  • NLP
  • OpenMP
  • S2S Compilers
  • Shared Memory Parallelism
  • Transformers

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Advising OpenMP Parallelization via A Graph-Based Approach with Transformers'. Together they form a unique fingerprint.

Cite this