Kenji Ono
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Kenji Ono has a total of 160 co-authors. Here are the top co-authors by number of shared publications:
Gaku Tanaka
Graduate School of Engineering, Chiba University
11 shared publicationsHideo Yokota
Image Processing Research Team, RIKEN
10 shared publicationsToshihiro Sera
Graduate School of Engineering, Kyusyu University
9 shared publications- Jorji Nonaka
RIKEN R-CCS, Kobe, Japan
8 shared publications Shu Takagi
Biomechanics Simulation Research Team, RIKEN
8 shared publicationsGaku Hashimoto
The University of Tokyo
6 shared publications
Latest Publications
- Open Access
- 12 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Large-scaleWGS Analysis on the supercomputer Fugaku
The remarkable improvement in the performance of next-generation sequencers has led to the analysis of large numbers of samples in recent years for whole-genome analysis. The computing capabilities required for the same have also increased dramatically; in particular, supercomputers and cloud services have become essential. However, few analyses have used TOP500 large-scale parallel computers because of the difference in the computational scientific properties of existing software for supercomputers and software for bioinformatics and the associated supercomputer resource management methods. In this study, we investigated the difference between bioinformatics pipelines and legacy supercomputer applications carefully and found that job management systems of TOP500-class supercomputers and distributed file systems have issues for bioinformatics pipelines. Based on the results, we developed some easy and efficient methods to overcome them, then we ported and optimized Genomon, a whole-genome analysis pipeline to Fugaku. We performed large-scale WGS analysis on Fugaku, and successfully analyzed over 1 thousand samples with 1,460 thousand node-hours, which is currently the largest analysis of the world’s. Availability: The ported and optimized Genomon is available at the following URL. https://bitbucket.org/genomon_wg/genomonpipeline/src/GPF_develop/
- 12 January 2024
- Published by Association for Computing Machinery (ACM)
- p. 14-20
Open AccessLarge-scaleWGS Analysis on the supercomputer Fugaku
The remarkable improvement in the performance of next-generation sequencers has led to the analysis of large numbers of samples in recent years for whole-genome analysis. The computing capabilities required for the same have also increased dramatically; in particular, supercomputers and cloud services have become essential. However, few analyses have used TOP500 large-scale parallel computers because of the difference in the computational scientific properties of existing software for supercomputers and software for bioinformatics and the associated supercomputer resource management methods. In this study, we investigated the difference between bioinformatics pipelines and legacy supercomputer applications carefully and found that job management systems of TOP500-class supercomputers and distributed file systems have issues for bioinformatics pipelines. Based on the results, we developed some easy and efficient methods to overcome them, then we ported and optimized Genomon, a whole-genome analysis pipeline to Fugaku. We performed large-scale WGS analysis on Fugaku, and successfully analyzed over 1 thousand samples with 1,460 thousand node-hours, which is currently the largest analysis of the world’s. Availability: The ported and optimized Genomon is available at the following URL. https://bitbucket.org/genomon_wg/genomonpipeline/src/GPF_develop/
- 5 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3299-3304
Acceleration of BAM I/O on distributed file systems
Rapid advances in high-throughput sequencers have made it possible to obtain large amounts of whole genome data quickly and inexpensively. As the amount of data increases, the increase in computation time has become a serious problem. One of the main causes of this problem is file I/O performance. Most pipelines do not implement file I/O suitable for distributed file systems, which are common storage systems in parallel computers used for large-scale analysis. In this study, we developed an I/O system adapted to distributed file systems. In the proposed system, storage access frequency was suppressed by I/O buffers, and thread parallelization by OpenMP was employed. We tested the developed system and a high parallel speedup was achieved. Availability: The developed system is freely available from https://github.com/SatoshiITO/GAPS
- 5 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 3299-3304
Acceleration of BAM I/O on distributed file systems
Rapid advances in high-throughput sequencers have made it possible to obtain large amounts of whole genome data quickly and inexpensively. As the amount of data increases, the increase in computation time has become a serious problem. One of the main causes of this problem is file I/O performance. Most pipelines do not implement file I/O suitable for distributed file systems, which are common storage systems in parallel computers used for large-scale analysis. In this study, we developed an I/O system adapted to distributed file systems. In the proposed system, storage access frequency was suppressed by I/O buffers, and thread parallelization by OpenMP was employed. We tested the developed system and a high parallel speedup was achieved. Availability: The developed system is freely available from https://github.com/SatoshiITO/GAPS
- 9 November 2022
- Published by Springer Nature in Journal of Visualization
- Vol. 26 (3) , 649-665
ESRGAN-based visualization for large-scale volume data
- 9 November 2022
- Published by Springer Nature in Journal of Visualization
- Vol. 26 (3) , 649-665
ESRGAN-based visualization for large-scale volume data
- 17 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 29 (12) , 5434-5450
Toward Feature-Preserving Vector Field Compression
The objective of this work is to develop error-bounded lossy compression methods to preserve topological features in 2D and 3D vector fields. Specifically, we explore the preservation of critical points in piecewise linear and bilinear vector fields. We define the preservation of critical points as, without any false positive, false negative, or false type in the decompressed data, (1) keeping each critical point in its original cell and (2) retaining the type of each critical point (e.g., saddle and attracting node). The key to our method is to adapt a vertex-wise error bound for each grid point and to compress input data together with the error bound field using a modified lossy compressor. Our compression algorithm can be also embarrassingly parallelized for large data handling and in situ processing. We benchmark our method by comparing it with existing lossy compressors in terms of false positive/negative/type rates, compression ratio, and various vector field visualizations with several scientific applications.
- 17 October 2022
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Visualization and Computer Graphics
- Vol. 29 (12) , 5434-5450
Toward Feature-Preserving Vector Field Compression
The objective of this work is to develop error-bounded lossy compression methods to preserve topological features in 2D and 3D vector fields. Specifically, we explore the preservation of critical points in piecewise linear and bilinear vector fields. We define the preservation of critical points as, without any false positive, false negative, or false type in the decompressed data, (1) keeping each critical point in its original cell and (2) retaining the type of each critical point (e.g., saddle and attracting node). The key to our method is to adapt a vertex-wise error bound for each grid point and to compress input data together with the error bound field using a modified lossy compressor. Our compression algorithm can be also embarrassingly parallelized for large data handling and in situ processing. We benchmark our method by comparing it with existing lossy compressors in terms of false positive/negative/type rates, compression ratio, and various vector field visualizations with several scientific applications.
- 1 October 2019
- Published by Springer Nature
- p. 116-122
Showing Ultra-High-Resolution Images in VDA-Based Scalable Displays
- 1 October 2019
- Published by Springer Nature
- p. 116-122
Showing Ultra-High-Resolution Images in VDA-Based Scalable Displays
- Open Access
- 3 February 2019
- Published by Wiley in Modelling and Simulation in Engineering
- Vol. 2019, 1-10
High-Performance Parallel Simulation of Airflow for Complex Terrain Surface
It is important to develop a reliable and high-throughput simulation method for predicting airflows in the installation planning phase of windmill power plants. This study proposes a two-stage mesh generation approach to reduce the meshing cost and introduces a hybrid parallelization scheme for atmospheric fluid simulations. The meshing approach splits mesh generation into two stages: in the first stage, the meshing parameters that uniquely determine the mesh distribution are extracted, and in the second stage, a mesh system is generated in parallel via an in situ approach using the parameters obtained in the initialization phase of the simulation. The proposed two-stage approach is flexible since an arbitrary number of processes can be selected at run time. An efficient OpenMP-MPI hybrid parallelization scheme using a middleware that provides a framework of parallel codes based on the domain decomposition method is also developed. The preliminary results of the meshing and computing performance show excellent scalability in the strong scaling test.
- 3 February 2019
- Published by Wiley in Modelling and Simulation in Engineering
- Vol. 2019, 1-10
Open AccessHigh-Performance Parallel Simulation of Airflow for Complex Terrain Surface
It is important to develop a reliable and high-throughput simulation method for predicting airflows in the installation planning phase of windmill power plants. This study proposes a two-stage mesh generation approach to reduce the meshing cost and introduces a hybrid parallelization scheme for atmospheric fluid simulations. The meshing approach splits mesh generation into two stages: in the first stage, the meshing parameters that uniquely determine the mesh distribution are extracted, and in the second stage, a mesh system is generated in parallel via an in situ approach using the parameters obtained in the initialization phase of the simulation. The proposed two-stage approach is flexible since an arbitrary number of processes can be selected at run time. An efficient OpenMP-MPI hybrid parallelization scheme using a middleware that provides a framework of parallel codes based on the domain decomposition method is also developed. The preliminary results of the meshing and computing performance show excellent scalability in the strong scaling test.
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2019.32, 068
Parallel particle tracking of 10 billion class in cyclone separator
In order to analyze the particle behavior in a cyclone separator in detail and predict the particle collection efficiency with high accuracy, we report the implementation of parallel computation of Euler-Lagrange coupling method that can handle massive particles. We implemented particle parallel computation based on the domain decomposition method to obtain the basic performance data and performed performance evaluation focusing on migration process. As a result, it was found that the migration process proposed has the property of asymptotically approaching constant cost without failure even if the number of particles increases.
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2019.32, 068
Parallel particle tracking of 10 billion class in cyclone separator
In order to analyze the particle behavior in a cyclone separator in detail and predict the particle collection efficiency with high accuracy, we report the implementation of parallel computation of Euler-Lagrange coupling method that can handle massive particles. We implemented particle parallel computation based on the domain decomposition method to obtain the basic performance data and performed performance evaluation focusing on migration process. As a result, it was found that the migration process proposed has the property of asymptotically approaching constant cost without failure even if the number of particles increases.
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2019.32, 211
Application of Parallel-in-Time Integration to the Phase-Field Models
The phase-field method has many applications because of the applicability of that method. Also, the phase field equation is one of parabolic PDEs, then that is useful for solving parabolic PDEs. In addition, simulation of many cases with the short time is expected from progress of the data science. However, there is no report of an evaluation of performance of parareal method to phase field equation. Therefore, in this study, we apply the parareal method to a 1D phase field equation and show the convergence data to consider the efficient use of parareal method. The results of this study show the following. The parareal method works for the phase field equation. The number of the parareal iteration is almost 10 for the number of time slices 100. The number of parareal iteration does not increase even if we increase the time coarsening ratio to 400. Backwarddifference 4th method enables an accurate convergence.
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of The Computational Mechanics Conference
- Vol. 2019.32, 211
Application of Parallel-in-Time Integration to the Phase-Field Models
The phase-field method has many applications because of the applicability of that method. Also, the phase field equation is one of parabolic PDEs, then that is useful for solving parabolic PDEs. In addition, simulation of many cases with the short time is expected from progress of the data science. However, there is no report of an evaluation of performance of parareal method to phase field equation. Therefore, in this study, we apply the parareal method to a 1D phase field equation and show the convergence data to consider the efficient use of parareal method. The results of this study show the following. The parareal method works for the phase field equation. The number of the parareal iteration is almost 10 for the number of time slices 100. The number of parareal iteration does not increase even if we increase the time coarsening ratio to 400. Backwarddifference 4th method enables an accurate convergence.
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of the Bioengineering Conference Annual Meeting of BED/JSME
- Vol. 2019.31, 1G22
Voxel-based simulation of nasal airflow during a sniff
- 1 January 2019
- Published by Japan Society of Mechanical Engineers in The Proceedings of the Bioengineering Conference Annual Meeting of BED/JSME
- Vol. 2019.31, 1G22
Voxel-based simulation of nasal airflow during a sniff