Stencil Computation has long been an omnipresent kernel of a wide range of scientific and engineering applications. There is much work investigating the stencil performance on x86 processors and accelerators such as G...
详细信息
Stencil Computation has long been an omnipresent kernel of a wide range of scientific and engineering applications. There is much work investigating the stencil performance on x86 processors and accelerators such as GPU. Meanwhile, ARM processors for HPC have been highlighted more and more with Fugaku of Japan becoming the first number one system on the TOP-500 list. In this paper, we focus on modeling and optimizing the performance of stencil computation on ARM architectures. Specifically, we proposed a performance model for stencil computation based on tiling optimization to guide the optimal configuration of tiling parameters for good cache reuse. We validate the proposed model with the Performance Monitor Unit (PMU) provided by the ARM processor. Experimental results show that the prediction error of the execution time can be lower to 1.37%. Furthermore, we can achieve a maximum of 1.26 × speedup compared to the naive implementation according to the optimization parameters provided from the model.
In this paper, we proposed image based indoor automatic flight system of quadcopter to follow a user defined flight guideline on the floor by utilizing its own camera installed in the bottom of the drone for autonomou...
详细信息
ISBN:
(纸本)9788993215182
In this paper, we proposed image based indoor automatic flight system of quadcopter to follow a user defined flight guideline on the floor by utilizing its own camera installed in the bottom of the drone for autonomous indoor flight of the quadcopter system. For this paper, we configure the hardware with quadcopter and GCS using notebook. GCS analyzes the sensor signals including image signal from quadcopter and send motor command signal to quadcopter. The autonomous flight algorithm is divided into an image processing system and a drone control system. This system analyzed the image signal from drone camera, and estimated direction and position of the guideline, and sent motor commands to drone. We verified the image based automatic flight control system by modeling and simulation. We confirmed that the quadcopter was autonomous flying properly along the guideline on the floor by modeling and simulation.
The complexity of modern automotive systems is mainly due to its rising electronic and software portions. In premium cars, more than one hundred Electronic Control Units (ECUs) are connected within a complex network t...
详细信息
ISBN:
(纸本)9781538683965
The complexity of modern automotive systems is mainly due to its rising electronic and software portions. In premium cars, more than one hundred Electronic Control Units (ECUs) are connected within a complex network to realize a wide variety of tasks ranging from infotainment and navigation to safety-critical functions such as driving assistance systems. In addition, the immense number of existing system variants and configurations makes the development, testing and integration of software a real challenge. A simulation environment that allows to test different system variants and configurations of automotive embeddedsystems is presented in this work. The considered proof of concept is a network of multiple ECUs with appropriate sensors and actuators realizing the function of an Adaptive Cruise Control (ACC). The virtual test environment, which has been implemented using Matlab/Simulink, includes a simulation of the vehicle dynamics and the external environment of the host car in form of a velocity profile of the lead vehicle. Further, it allows to configure, load and test different configurations of the system using a convenient variant management method. Besides, a virtual platform can be co-simulated allowing a virtual hardware-in-the-loop (vHiL) simulation which encloses the target C-code of one or more ECUs.
The Food and Agriculture Organization (FAO) estimates that if female farmers in developing countries had access to the same resources as men, the number of undernourished people would decrease by 12%-17% [9]. Clearly,...
详细信息
ISBN:
(数字)9783030217419
ISBN:
(纸本)9783030217419;9783030217402
The Food and Agriculture Organization (FAO) estimates that if female farmers in developing countries had access to the same resources as men, the number of undernourished people would decrease by 12%-17% [9]. Clearly, gender equity is a vital part of increasing agricultural production to feed the world's projected 9.7 billion people by 2050. However, programs designed to empower women in agricultural systems are expensive, and no quantitative model exists to define and explore the efficacy of policies in cultural contexts. We introduce a formal model of household decisions embedded in an agent-based model of community gender dynamics and show how the explicit definition of gender inequity can help inform decision-making about programs intended to empower women.
This paper studies AI modeling for the solder joint fatigue risk estimation under the thermal cycle loading of redistributed wafer level packaging. The artificial neural network (ANN), recurrent neural network (RNN) a...
详细信息
ISBN:
(数字)9781728160498
ISBN:
(纸本)9781728160504
This paper studies AI modeling for the solder joint fatigue risk estimation under the thermal cycle loading of redistributed wafer level packaging. The artificial neural network (ANN), recurrent neural network (RNN) and vectorized-gate network long short-term memory (VNLSTM) architectures have been trained by the same dataset to investigate their performance for this task. The learning accuracy criterion, the implementation of all neural network architecture, the learning results and result analysis would be *** the involvement of the time/temperature-dependent nonlinearity material characteristics, it is recommended that more than three hidden layers and a proper neural network architecture, which is capable of the sequential data processing, should be considered in order to guarantee the required accuracy and the satisfied convergence speed.
The work is devoted to research in the direction of creating multi-position systems based on low-power transceivers for the construction of inter-satellite communication channels of the terahertz range of low-orbit sa...
详细信息
The work is devoted to research in the direction of creating multi-position systems based on low-power transceivers for the construction of inter-satellite communication channels of the terahertz range of low-orbit satellite systems with a “distributed satellite” architecture. Comprehensive studies of the possibilities of creating communication channels of the terahertz range for communication systems based on high-altitude air platforms are being carried out. Models for improving the efficiency of terahertz radio communication systems are proposed: an improved algorithm for searching for multibeam signals in radio systems based on high-altitude THz band airplanes and a simulation model for predicting the maximum length of the terahertz range wireless communication system U-IR-based. The use of X-ray diffraction in an advanced algorithm has simplified the implementation of multi-beam signal processing and significantly reduced the computational processing resource. The model on the basis of machine learning allows in laboratory conditions to investigate experimentally various architectures of construction of devices and systems of generation 5G/IoT and to work out knots of formation and processing of information signals and to define their best characteristics.
The proceedings contain 29 papers. The special focus in this conference is on computer Networks. The topics include: Cyber Attack Detection by Using Neural Network Approaches: Shallow Neural Network, Deep Neural Netwo...
ISBN:
(纸本)9783030219512
The proceedings contain 29 papers. The special focus in this conference is on computer Networks. The topics include: Cyber Attack Detection by Using Neural Network Approaches: Shallow Neural Network, Deep Neural Network and AutoEncoder;temporal Characteristics of CodeSys Programmed Raspberry Pi and Beaglebone Black embedded Devices;ateb-modeling Method Application for High Quality Video Traffic modeling;a Comparison of Request Distribution Strategies Used in One and Two Layer architectures of Web Cloud systems;mobile Voice Traffic Load Characteristics;vehicle in Motion Weighing Based on Vibration Data Collected from Sensor Network;Optimal Spectrum Sharing in RF Energy Harvesting Cognitive Femtocell Networks;Accuracy of UWB Location Tracking Devices When on the Move;cells Interrelation in Mobile Networks;an Architectural Framework Proposal for IoT Driven Agriculture;communication Model of Smart Substation for Cyber-Detection systems;virtual Oversampling of Sparse Channel Impulse Response;recreation of Containers for High Availability Architecture and Container-Based Applications;analysis of Energy Consumption in Cloud Center with Tasks Migrations;queueing systems with Non-homogeneous Customers and Infinite Sectorized Memory Space;transient Solution of a Heterogeneous Queuing System with Balking and Retention of Reneging Customers;Cluster-Based Web System Models for Different Classes of Clients in QPN;A Queueing Model and Performance Analysis of UPnP/HTTP Client Server Interactions in Networked Control systems;a System with Warm Standby;AQM Mechanism with the Dropping Packet Function Based on the Answer of Several PIα Controllers;the Procedure of Key Distribution in Military IoT Networks;modeling of computer Networks Using SAP HANA Smart Data Streaming.
This paper presents a novel ferroelectric field-effect transistor (FeFET) in-memory computing architecture dedicated to accelerate Binary Neural Networks (BNNs). We present in-memory convolution, batch normalization a...
详细信息
ISBN:
(数字)9781728171470
ISBN:
(纸本)9781728172798
This paper presents a novel ferroelectric field-effect transistor (FeFET) in-memory computing architecture dedicated to accelerate Binary Neural Networks (BNNs). We present in-memory convolution, batch normalization and dense layer processing through a grid of small crossbars with reduced unit size, which enables multiple bit operation and value accumulation. Additionally, we explore the possible operations parallelization for maximized computational performance. simulation results show that our new architecture achieves a computing performance up to 2.46 TOPS while achieving a high power efficiency reaching 111.8 TOPS/Watt and an area of 0.026 mm 2 in 22nm FDSOI technology.
In the last decades, vehicular control applications evolved to comprehensive Advanced Driver Assistance systems (ADAS) with exploding complexity. Several simulation-driven approaches can ease their development, e.g., ...
详细信息
The great progress in the field of digitization and networking of the industry brings many advantages as well as new challenges. On the one hand, it enables new business models that hold great economic growth potentia...
详细信息
ISBN:
(数字)9781728172194
ISBN:
(纸本)9781728172200
The great progress in the field of digitization and networking of the industry brings many advantages as well as new challenges. On the one hand, it enables new business models that hold great economic growth potential. On the other hand, it leads to an increased heterogeneity of the resulting value- added networks. This poses a striking problem when it comes to the interconnection of industrial devices and the establishment of new business models. Since the disadvantages can quickly outweigh the advantages, activities in this area proceed still very cautiously. A solution is offered by highly dynamic attribute-based usage control architectures at network-level, which monitor the entire communication in the value-added networks. Previous approaches, however, either have the disadvantage that they can only use communication metadata as attributes, or are highly specialized and tailored to the respective systems of the individual network nodes. In addition, they are usually pure access control systems instead of usage control systems. In this paper, the authors propose a novel attribute-based usage control model (ABUC) that allows simulation models in form of digital twins of the respective network components to be used in order to predict their behavior and thus control network traffic according to predicted usage. The ABUC model is then implemented in a use case and validated with several performance tests.
暂无评论