Pu Han

Energy distribution networks represent crucial infrastructures for modern society, and various simulation tools have been widely used by energy suppliers to manage these intricate networks. However, simulation calculations include a large number of fluid control equations, and computational overhead limits the performance of simulation software. This paper proposes a universal parallel simulation framework for energy pipeline networks that takes advantages of data parallelism and computational independence between network elements. A non-pipe model of an energy supply network is optimized, and the input and output of the network model in the proposed framework are modified, which can reduce the development burden during the numerical computations of the pipeline network and weaken the computational correlation between different simulated components. In addition, independent computations can be performed concurrently through periodic data exchange procedures between component instances, improving the parallelism and efficiency of simulation computations. Further, a parallel water pipelines network simulation computing paradigm based on a heterogeneous computer hardware architecture is used to evaluate the proposed framework’s performance. A series of tests are conducted to verify the accuracy of the proposed framework, and simulation errors of less than 5% are achieved. The results of multi-threaded simulation experiments have demonstrated the feasibility of the proposed framework in a parallel computing approach. Moreover, an Advanced Micro Devices (AMD) Deep Computing Unit (DCU)-parallel program is implemented into a water supply network simulation system; the computational efficiency of this system is compared with that of its serial counterpart. The experimental results show that the proposed framework is appropriate for high-performance computer architectures, and the 18x speed-up ratio demonstrates that the parallel program based on the proposed universal framework outperforms the serial program. That provides the basis for the application of pipe network simulation on high-performance computers.

In order to give full play to the advantages of DCU accelerator and solve the problems of algorithm SpMV (Sparse matrix-vector multiplication) with limited bandwidth, unbalanced load, and non-combined memory access, a SCSR (Static Compressed Sparse Row) using CSR (Compressed Sparse Row) storage format is proposed based on DCU accelerator. The algorithm statically allocates the same number of rows to each thread block according to the average number of non-zero elements in each row to avoid unnecessary computations; the application of storage resources is reduced by reusing the on-chip high-speed storage space LDS (Local Data Shared), thus improving the CU (Compute Unit) occupancy. The experiment uses 15 sparse matrices in different fields for testing. The results show that compared with the SpMV algorithm in the hipSPARSE library, the SCSR algorithm achieves an average speedup ratio of 4.83 times.

As an emerging computing model, edge computing greatly expands the collaboration capabilities of the servers. It makes full use of the available resources around the users to quickly complete the task request coming from the terminal devices. Task offloading is a common solution for improving the efficiency of task execution on edge networks. However, the peculiarities of the edge networks, especially the random access of mobile devices, brings unpredictable challenges to the task offloading in a mobile edge network. In this paper, we propose a trajectory prediction model for moving targets in edge networks without users’ historical paths which represents their habitual movement trajectory. We also put forward a mobility-aware parallelizable task offloading strategy based on a trajectory prediction model and parallel mechanisms of tasks. In our experiments, we compared the hit ratio of the prediction model, network bandwidth and task execution efficiency of the edge networks by using the EUA data set. Experimental results showed that our model is much better than random, non-position prediction parallel, non-parallel strategy-based position prediction. Where the task offloading hit rate is closed to the user’s moving speed, when the speed is less 12.96 m/s, the hit rate can reach more than 80%. Meanwhile, we we also find that the bandwidth occupancy is significantly related to the degree of task parallelism and the number of services running on servers in the network. The parallel strategy can boost network bandwidth utilization by more than eight times when compared to a non-parallel policy as the number of parallel activities grows.

SpMV(Sparse matrix-vector multiplication) plays an important role in the fields of scientific research and engineering, but its non-merged memory access and load imbalance cause worse performance when dealing with very large-scale sparse matrices. In view of the above problems, this paper proposes a SpMV algorithm of MPI+HIP hybrid model based on the DCU accelerator, which uses on-chip high-speed storage space at a fine-grained level to reduce the hidden access latency to global memory. Through the warp reduction, the barriers of implicit synchronization and thread synchronization logic are reduced. The MPI is used at the coarse-grained level to partition the ultra-large-scale sparse matrix in behavioral units, and computational tasks are assigned to processes on the corresponding DCUs for accelerated processing. Experimental results show that, compared with the SpMV algorithm in the hipSPARSE library, the SpMV algorithm of the hybrid programming model achieves an increase of speedup from 4.34 to 6.44, which significantly improves the parallel efficiency.

The localization of continuous objects and the scheduling of resources are challenging issues in wireless sensor networks (WSNs). Due to the irregular shape of the continuous target area and the sensor deployment in WSNs, the sensor data are always discrete and sparse, and most network resources are also limited by the node energy. To achieve faster detection and tracking of continuous objects, we propose a convolution-based continuous object localization algorithm (named CCOL). Moreover, we implement the idea of greedy and dynamic programming to design an energy-saving and efficient strategy model (named MSSM) to respond to emergencies caused by multiple continuous targets in most specific WSNs. The simulation experiments demonstrate that CCOL is superior to other localization algorithms in terms of time complexity and execution performance. Furthermore, the feasibility of the multinode scheduling strategy is verified by setting different mobile nodes to respond to the target area in certain green WSNs.

The application of deep neural networks (DNNs) for road extraction from remote sensing images has gained broad interest because of the competence concerning complex nonlinear relations; however, the presence of noisy labels in the training data sets adversely affects the performance of DNNs. The existing methods of improving the robustness of DNNs focus on modeling the noise distribution. However, these approaches are not satisfactory because of the inaccurate high-level image features obtained by the DNNs. To address this issue, we develop a noise probabilistic model for learning the label noise based on the relationship between the input images, noisy labels, and true labels. The key idea of the probabilistic model is to directly explore the information from the input images and apply it to model the label noise. Then, a robust deep neural network (RDNN) is proposed to instantiate the noise probabilistic model, which consists of two important modules: the true label predictor (TLP) and the noise label estimator (NLE). Especially, the TLP is made of a DNN with softmax, which is used to learn the true label distribution. The NLE is applied to model the label noise distribution, which aims to absorb the label noise in the training process. Moreover, to tackle the challenges in the optimization, we deduce a loss function with the novel regularization, which allows the RDNN to conduct effective training on the noise data set. The effectiveness of the proposed method is validated by experiments on three road data sets that contain various resolutions and imaging conditions. The results demonstrate its superiority over state-of-the-art methods in visual performance and classification accuracy.