OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation
OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation
Tal Kadosh, Niranjan Hasabnis, Prema Soundararajan, Vy A. Vo, Mihai Capotă, Nesreen K. Ahmed, Yuval Pinter, Gal Oren
Machine Learning for Systems, MLforSys
NeurIPS 2024
Existing automatic code parallelization tools are either too conservative (formal-methods-based tools) or too inaccurate (AI-based tools). This paper introduces OMPar, an AI-driven tool that breaks the problem into two sub-problems of parallelism detection and parallel pragma generation and then integrates two state-of-the-art models to solve the problem. We evaluate OMPar and competing existing tools in terms of accuracy (in suggesting correct pragma), syntax, semantics, and run-time performance of suggested pragmas. Overall, we found that OMPar outperforms existing tools in accurately suggesting parallelization pragmas. Moreover, we found that OMPar-suggested pragmas are also syntactically- and semantically valid (high compilation and test success rate), and they also deliver performance improvement over corresponding baseline serial programs. The sources of this work are available in our repository: https://github.com/Scientific-Computing-Lab/OMPar
Видео OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation канала Scientific Computing Lab
Tal Kadosh, Niranjan Hasabnis, Prema Soundararajan, Vy A. Vo, Mihai Capotă, Nesreen K. Ahmed, Yuval Pinter, Gal Oren
Machine Learning for Systems, MLforSys
NeurIPS 2024
Existing automatic code parallelization tools are either too conservative (formal-methods-based tools) or too inaccurate (AI-based tools). This paper introduces OMPar, an AI-driven tool that breaks the problem into two sub-problems of parallelism detection and parallel pragma generation and then integrates two state-of-the-art models to solve the problem. We evaluate OMPar and competing existing tools in terms of accuracy (in suggesting correct pragma), syntax, semantics, and run-time performance of suggested pragmas. Overall, we found that OMPar outperforms existing tools in accurately suggesting parallelization pragmas. Moreover, we found that OMPar-suggested pragmas are also syntactically- and semantically valid (high compilation and test success rate), and they also deliver performance improvement over corresponding baseline serial programs. The sources of this work are available in our repository: https://github.com/Scientific-Computing-Lab/OMPar
Видео OMPar: Automatic Parallelization with AI-Driven Source-to-Source Compilation канала Scientific Computing Lab
Комментарии отсутствуют
Информация о видео
10 декабря 2024 г. 0:21:00
00:08:47
Другие видео канала
Quantifying OpenMP: Statistical Insights into Usage and Adoption
CXL Memory as Persistent Memory for Disaggregated HPC (A Practical Approach)
Shared-Memory parallelism: CPUs, GPUs, and in-between - Lecture 8 (Yehonatan Fridman)
Shared-Memory Parallelism: CPUs, GPUs and in-between - Lecture 5
Compile and Run ScalSALE on IntelDevCloud
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 3
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 2
MPIrigen: MPI Code Generation through Domain-Specific Language Models (AI4Sys - HPDC'24)
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 11 (by Guy Tamir)
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 7
Shared-Memory Parallelism: CPUs, GPUs and in-between - Lecture 8
Shared-Memory Parallelism: CPUs, GPUs and in-between - Lecture 4
Shared-Memory Parallelism: CPUs, GPUs and in-between - Lecture 9
Avoiding Post-Processing with Context: Texture Boundary Detection in Metallography
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 9
Automatic Parallelization for Concurrent Programming - Invited Talk by Re'em Harel (Lecture 10)
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 10
Shared-Memory Parallelism: CPUs, GPUs and in-between - Lecture 3
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 12
Shared-Memory Parallelism: CPUs, GPUs and in-between - Practice 4
Open Software for the Parallel, Heterogeneous Future - Invited Talk by Guy Tamir (Lecture 11)
