Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model
Stencil computation is one kind of the most important loop kernels in scientific and engineering computing applications.Loop tiling can effectively improve the data locality of stencil computation and the degree of computational parallelism,but the best tile size is hard to choose.Traditional tile s...
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | zho |
| Published: |
Editorial office of Computer Science
2022-10-01
|
| Series: | Jisuanji kexue |
| Subjects: | |
| Online Access: | https://www.jsjkx.com/fileup/1002-137X/PDF/1002-137X-2022-49-10-18.pdf |
| _version_ | 1797845134995357696 |
|---|---|
| author | BAO Yi-kun, ZHANG Peng, XU Xiao-wen, MO Ze-yao |
| author_facet | BAO Yi-kun, ZHANG Peng, XU Xiao-wen, MO Ze-yao |
| author_sort | BAO Yi-kun, ZHANG Peng, XU Xiao-wen, MO Ze-yao |
| collection | DOAJ |
| description | Stencil computation is one kind of the most important loop kernels in scientific and engineering computing applications.Loop tiling can effectively improve the data locality of stencil computation and the degree of computational parallelism,but the best tile size is hard to choose.Traditional tile size selection methods usually have shortcomings in some ways of time overhead,labor cost and model accuracy.In this paper,a tile size selection method based on artificial neural network is proposed to predict the optimal tile size of three-dimensional Jacobi stencil loop programs.Experimental results show that,for 11 real stencil programs,the performance improvement of the programs using the model prediction tile size compared with the non tiling is 2% and 35% in serial and parallel tests respectively.Compared with the well-known grid search method,our method has a similar prediction accuracy,but only takes one 30 thousandth of the online time cost.In addition,compared with the Turbo-tiling method,our method improves the performance of tiled codes nearly 9% in average. |
| first_indexed | 2024-04-09T17:33:41Z |
| format | Article |
| id | doaj.art-7674ba461cec4bc6900af00ea21ae6ad |
| institution | Directory Open Access Journal |
| issn | 1002-137X |
| language | zho |
| last_indexed | 2024-04-09T17:33:41Z |
| publishDate | 2022-10-01 |
| publisher | Editorial office of Computer Science |
| record_format | Article |
| series | Jisuanji kexue |
| spelling | doaj.art-7674ba461cec4bc6900af00ea21ae6ad2023-04-18T02:32:39ZzhoEditorial office of Computer ScienceJisuanji kexue1002-137X2022-10-014910182610.11896/jsjkx.220100147Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network ModelBAO Yi-kun, ZHANG Peng, XU Xiao-wen, MO Ze-yao01 Graduate School of China Academy of Engineering Physics,Beijing 100094,China ;2 Institute of Applied Physics and Computational Mathematics,Beijing 100088,China ;3 CAEP Software Center for High Performance Numerical Simulation,Beijing 100088,China ;4 China Academy of Engineering Physics,Mianyang,Sichuan 621900,ChinaStencil computation is one kind of the most important loop kernels in scientific and engineering computing applications.Loop tiling can effectively improve the data locality of stencil computation and the degree of computational parallelism,but the best tile size is hard to choose.Traditional tile size selection methods usually have shortcomings in some ways of time overhead,labor cost and model accuracy.In this paper,a tile size selection method based on artificial neural network is proposed to predict the optimal tile size of three-dimensional Jacobi stencil loop programs.Experimental results show that,for 11 real stencil programs,the performance improvement of the programs using the model prediction tile size compared with the non tiling is 2% and 35% in serial and parallel tests respectively.Compared with the well-known grid search method,our method has a similar prediction accuracy,but only takes one 30 thousandth of the online time cost.In addition,compared with the Turbo-tiling method,our method improves the performance of tiled codes nearly 9% in average.https://www.jsjkx.com/fileup/1002-137X/PDF/1002-137X-2022-49-10-18.pdfstencil computation|loop tiling technology|machine learning|artificial neural network |
| spellingShingle | BAO Yi-kun, ZHANG Peng, XU Xiao-wen, MO Ze-yao Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model Jisuanji kexue stencil computation|loop tiling technology|machine learning|artificial neural network |
| title | Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model |
| title_full | Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model |
| title_fullStr | Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model |
| title_full_unstemmed | Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model |
| title_short | Prediction of Optimal Loop Tiling Size for stencil Computation Based on Neural Network Model |
| title_sort | prediction of optimal loop tiling size for stencil computation based on neural network model |
| topic | stencil computation|loop tiling technology|machine learning|artificial neural network |
| url | https://www.jsjkx.com/fileup/1002-137X/PDF/1002-137X-2022-49-10-18.pdf |
| work_keys_str_mv | AT baoyikunzhangpengxuxiaowenmozeyao predictionofoptimallooptilingsizeforstencilcomputationbasedonneuralnetworkmodel |