Conventional gradient-based full waveform inversion (FWI) is a local optimization, which is highly dependent on the initial model and prone to trapping in local minima. Globally optimal FWI that can overcome this limi...Conventional gradient-based full waveform inversion (FWI) is a local optimization, which is highly dependent on the initial model and prone to trapping in local minima. Globally optimal FWI that can overcome this limitation is particularly attractive, but is currently limited by the huge amount of calculation. In this paper, we propose a globally optimal FWI framework based on GPU parallel computing, which greatly improves the efficiency, and is expected to make globally optimal FWI more widely used. In this framework, we simplify and recombine the model parameters, and optimize the model iteratively. Each iteration contains hundreds of individuals, each individual is independent of the other, and each individual contains forward modeling and cost function calculation. The framework is suitable for a variety of globally optimal algorithms, and we test the framework with particle swarm optimization algorithm for example. Both the synthetic and field examples achieve good results, indicating the effectiveness of the framework. .展开更多
Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes...Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes the performance gains from parallel versus sequential hyperparameter optimization. Using scikit-learn’s Randomized SearchCV, this project tuned a Random Forest classifier for fake news detection via randomized grid search. Setting n_jobs to -1 enabled full parallelization across CPU cores. Results show the parallel implementation achieved over 5× faster CPU times and 3× faster total run times compared to sequential tuning. However, test accuracy slightly dropped from 99.26% sequentially to 99.15% with parallelism, indicating a trade-off between evaluation efficiency and model performance. Still, the significant computational gains allow more extensive hyperparameter exploration within reasonable timeframes, outweighing the small accuracy decrease. Further analysis could better quantify this trade-off across different models, tuning techniques, tasks, and hardware.展开更多
针对大数据环境下并行深度森林算法中存在不相关及冗余特征过多、多粒度扫描不平衡、分类性能不足以及并行化效率低等问题,提出了基于互信息和融合加权的并行深度森林算法(parallel deep forest algorithm based on mutual information ...针对大数据环境下并行深度森林算法中存在不相关及冗余特征过多、多粒度扫描不平衡、分类性能不足以及并行化效率低等问题,提出了基于互信息和融合加权的并行深度森林算法(parallel deep forest algorithm based on mutual information and mixed weighting,PDF-MIMW)。首先,在特征降维阶段提出了基于互信息的特征提取策略(feature extraction strategy based on mutual information,FE-MI),结合特征重要性、交互性和冗余性度量过滤原始特征,剔除过多的不相关和冗余特征;接着,在多粒度扫描阶段提出了基于填充的改进多粒度扫描策略(improved multi-granularity scanning strategy based on padding,IMGS-P),对精简后的特征进行填充并对窗口扫描后的子序列进行随机采样,保证多粒度扫描的平衡;其次,在级联森林构建阶段提出了并行子森林构建策略(sub-forest construction strategy based on mixed weighting,SFC-MW),结合Spark框架并行构建加权子森林,提升模型的分类性能;最后,在类向量合并阶段提出基于混合粒子群算法的负载均衡策略(load balancing strategy based on hybrid particle swarm optimization algorithm,LB-HPSO),优化Spark框架中任务节点的负载分配,降低类向量合并时的等待时长,提高模型的并行化效率。实验表明,PDF-MIMW算法的分类效果更佳,同时在大数据环境下的训练效率更高。展开更多
To decrease the time of generating a closure, a parallel algorithm of generating the closure of a resource description framework schema （RDFS） source is presented. In the algorithm, RDFS triples in the source are cl...To decrease the time of generating a closure, a parallel algorithm of generating the closure of a resource description framework schema （RDFS） source is presented. In the algorithm, RDFS triples in the source are classified according to the forms of triples in the entailment rules and it reduces the scope of searching for specific triples. The dependence among the classes of triples is analyzed. Based on the classification, the initial RDFS source is partitioned into several subsets. The subsets are distributed to each process, and the closure is generated in parallel by applying the RDFS entailment rules. Parallel generating the closure of an RDFS source takes less time and increases efficiency.展开更多
In this paper, a mathematical model consisting of forward and backward models is built on parallel genetic algorithms (PGAs) for fault diagnosis in a transmission power system. A new method to reduce the scale of faul...In this paper, a mathematical model consisting of forward and backward models is built on parallel genetic algorithms (PGAs) for fault diagnosis in a transmission power system. A new method to reduce the scale of fault sections is developed in the forward model and the message passing interface (MPI) approach is chosen to parallel the genetic algorithms by global sin-gle-population master-slave method (GPGAs). The proposed approach is applied to a sample system consisting of 28 sections, 84 protective relays and 40 circuit breakers. Simulation results show that the new model based on GPGAs can achieve very fast computation in online applications of large-scale power systems.展开更多
The desire to increase spatial and temporal resolution in modeling groundwater system has led to the requirement for intensive computational ability and large memory space. In the course of satisfying such requirement...The desire to increase spatial and temporal resolution in modeling groundwater system has led to the requirement for intensive computational ability and large memory space. In the course of satisfying such requirement, parallel computing has played a core role over the past several decades. This paper reviews the parallel algebraic linear solution methods and the parallel implementation technologies for groundwater simulation. This work is carried out to provide guidance to enable modelers of groundwater systems to make sensible choices when developing solution methods based upon the current state of knowledge in parallel computing.展开更多
Peta-scale high-perfomlance computing systems are increasingly built with heterogeneous CPU and GPU nodes to achieve higher power efficiency and computation throughput. While providing unprecedented capabilities to co...Peta-scale high-perfomlance computing systems are increasingly built with heterogeneous CPU and GPU nodes to achieve higher power efficiency and computation throughput. While providing unprecedented capabilities to conduct computational experiments of historic significance, these systems are presently difficult to program. The users, who are domain experts rather than computer experts, prefer to use programming models closer to their domains （e.g., physics and biology） rather than MPI and OpenME This has led the development of domain-specific programming that provides domain-specific programming interfaces but abstracts away some performance-critical architecture details. Based on experience in designing large-scale computing systems, a hybrid programming framework for scientific computing on heterogeneous architectures is proposed in this work. Its design philosophy is to provide a collaborative mechanism for domain experts and computer experts so that both domain-specific knowledge and performance-critical architecture details can be adequately exploited. Two real-world scientific applications have been evaluated on TH-IA, a peta-scale CPU-GPU heterogeneous system that is currently the 5th fastest supercomputer in the world. The experimental results show that the proposed framework is well suited for developing large-scale scientific computing applications on peta-scale heterogeneous CPU/GPU systems.展开更多
Recently, linear motors can have high speed control, high acceleration-deceleration. So linear motors are widely used in industrial applications such as precision machine tools. In our laboratory focusing on transport...Recently, linear motors can have high speed control, high acceleration-deceleration. So linear motors are widely used in industrial applications such as precision machine tools. In our laboratory focusing on transport system, we propose parallel synchronous drive of used the PM-LSM （permanent magnet linear synchronous motor）. It can pass luggage without having to stop the working. When you establish ＂parallel synchronous drive＂, a motor follows the other motor. In our laboratory, one of the motors is called ＂master motor＂ and the other motor called ＂slave motor＂. The master motor＇s speed and position pass slave motor then establish parallel synchronous drive. Therefore, slave motor requires high-responsive and precision that follows the master motor. This paper focuses on the control of the slave motor.展开更多
文摘Conventional gradient-based full waveform inversion (FWI) is a local optimization, which is highly dependent on the initial model and prone to trapping in local minima. Globally optimal FWI that can overcome this limitation is particularly attractive, but is currently limited by the huge amount of calculation. In this paper, we propose a globally optimal FWI framework based on GPU parallel computing, which greatly improves the efficiency, and is expected to make globally optimal FWI more widely used. In this framework, we simplify and recombine the model parameters, and optimize the model iteratively. Each iteration contains hundreds of individuals, each individual is independent of the other, and each individual contains forward modeling and cost function calculation. The framework is suitable for a variety of globally optimal algorithms, and we test the framework with particle swarm optimization algorithm for example. Both the synthetic and field examples achieve good results, indicating the effectiveness of the framework. .
文摘Hyperparameter tuning is a key step in developing high-performing machine learning models, but searching large hyperparameter spaces requires extensive computation using standard sequential methods. This work analyzes the performance gains from parallel versus sequential hyperparameter optimization. Using scikit-learn’s Randomized SearchCV, this project tuned a Random Forest classifier for fake news detection via randomized grid search. Setting n_jobs to -1 enabled full parallelization across CPU cores. Results show the parallel implementation achieved over 5× faster CPU times and 3× faster total run times compared to sequential tuning. However, test accuracy slightly dropped from 99.26% sequentially to 99.15% with parallelism, indicating a trade-off between evaluation efficiency and model performance. Still, the significant computational gains allow more extensive hyperparameter exploration within reasonable timeframes, outweighing the small accuracy decrease. Further analysis could better quantify this trade-off across different models, tuning techniques, tasks, and hardware.
文摘针对大数据环境下并行深度森林算法中存在不相关及冗余特征过多、多粒度扫描不平衡、分类性能不足以及并行化效率低等问题,提出了基于互信息和融合加权的并行深度森林算法(parallel deep forest algorithm based on mutual information and mixed weighting,PDF-MIMW)。首先,在特征降维阶段提出了基于互信息的特征提取策略(feature extraction strategy based on mutual information,FE-MI),结合特征重要性、交互性和冗余性度量过滤原始特征,剔除过多的不相关和冗余特征;接着,在多粒度扫描阶段提出了基于填充的改进多粒度扫描策略(improved multi-granularity scanning strategy based on padding,IMGS-P),对精简后的特征进行填充并对窗口扫描后的子序列进行随机采样,保证多粒度扫描的平衡;其次,在级联森林构建阶段提出了并行子森林构建策略(sub-forest construction strategy based on mixed weighting,SFC-MW),结合Spark框架并行构建加权子森林,提升模型的分类性能;最后,在类向量合并阶段提出基于混合粒子群算法的负载均衡策略(load balancing strategy based on hybrid particle swarm optimization algorithm,LB-HPSO),优化Spark框架中任务节点的负载分配,降低类向量合并时的等待时长,提高模型的并行化效率。实验表明,PDF-MIMW算法的分类效果更佳,同时在大数据环境下的训练效率更高。
基金The Weaponry Equipment Foundation of PLA Equipment Ministry (No.51406020105JB8103).
文摘To decrease the time of generating a closure, a parallel algorithm of generating the closure of a resource description framework schema （RDFS） source is presented. In the algorithm, RDFS triples in the source are classified according to the forms of triples in the entailment rules and it reduces the scope of searching for specific triples. The dependence among the classes of triples is analyzed. Based on the classification, the initial RDFS source is partitioned into several subsets. The subsets are distributed to each process, and the closure is generated in parallel by applying the RDFS entailment rules. Parallel generating the closure of an RDFS source takes less time and increases efficiency.
基金the National Natural Science Foundation of China (No. 50677062)the New Century Excellent Talents in Uni-versity of China (No. NCET-07-0745)the Natural Science Foundation of Zhejiang Province, China (No. R107062)
文摘In this paper, a mathematical model consisting of forward and backward models is built on parallel genetic algorithms (PGAs) for fault diagnosis in a transmission power system. A new method to reduce the scale of fault sections is developed in the forward model and the message passing interface (MPI) approach is chosen to parallel the genetic algorithms by global sin-gle-population master-slave method (GPGAs). The proposed approach is applied to a sample system consisting of 28 sections, 84 protective relays and 40 circuit breakers. Simulation results show that the new model based on GPGAs can achieve very fast computation in online applications of large-scale power systems.
基金supported by the National Basic Research Program (973 Program) of China under Grant No.2010CB428804 and 2011CB 309702
文摘The desire to increase spatial and temporal resolution in modeling groundwater system has led to the requirement for intensive computational ability and large memory space. In the course of satisfying such requirement, parallel computing has played a core role over the past several decades. This paper reviews the parallel algebraic linear solution methods and the parallel implementation technologies for groundwater simulation. This work is carried out to provide guidance to enable modelers of groundwater systems to make sensible choices when developing solution methods based upon the current state of knowledge in parallel computing.
基金Project(61170049) supported by the National Natural Science Foundation of ChinaProject(2012AA010903) supported by the National High Technology Research and Development Program of China
文摘Peta-scale high-perfomlance computing systems are increasingly built with heterogeneous CPU and GPU nodes to achieve higher power efficiency and computation throughput. While providing unprecedented capabilities to conduct computational experiments of historic significance, these systems are presently difficult to program. The users, who are domain experts rather than computer experts, prefer to use programming models closer to their domains （e.g., physics and biology） rather than MPI and OpenME This has led the development of domain-specific programming that provides domain-specific programming interfaces but abstracts away some performance-critical architecture details. Based on experience in designing large-scale computing systems, a hybrid programming framework for scientific computing on heterogeneous architectures is proposed in this work. Its design philosophy is to provide a collaborative mechanism for domain experts and computer experts so that both domain-specific knowledge and performance-critical architecture details can be adequately exploited. Two real-world scientific applications have been evaluated on TH-IA, a peta-scale CPU-GPU heterogeneous system that is currently the 5th fastest supercomputer in the world. The experimental results show that the proposed framework is well suited for developing large-scale scientific computing applications on peta-scale heterogeneous CPU/GPU systems.
文摘Recently, linear motors can have high speed control, high acceleration-deceleration. So linear motors are widely used in industrial applications such as precision machine tools. In our laboratory focusing on transport system, we propose parallel synchronous drive of used the PM-LSM （permanent magnet linear synchronous motor）. It can pass luggage without having to stop the working. When you establish ＂parallel synchronous drive＂, a motor follows the other motor. In our laboratory, one of the motors is called ＂master motor＂ and the other motor called ＂slave motor＂. The master motor＇s speed and position pass slave motor then establish parallel synchronous drive. Therefore, slave motor requires high-responsive and precision that follows the master motor. This paper focuses on the control of the slave motor.
