In parallel with the development of data transmission in the telecommunications sector to connect peripheral devices, hardware engineers have defined interfaces for independent communication systems. The basic idea of...
详细信息
In parallel with the development of data transmission in the telecommunications sector to connect peripheral devices, hardware engineers have defined interfaces for independent communication systems. The basic idea of having standardized interfaces for external devices was quickly extended to control and instrumentation equipment, as the real-time characteristic of distributedsystems became a challenge in the field. The proposed Modbus Extension (ModbusE) builds on these new techniques introduced in the UART interface, so that the newly introduced extension retains all the characteristics of the Modbus RTU except for the character bit structure. The validation of the ModbusE protocol allowed the proposal of a new communication message structure, a description of Modbus devices, an acquisition cycle to attain a deterministic temporal response behavior and the definition of an architecture for IIoT integration. In this work, practical research is based on data update times in SMARTConvert software when between 1 and 50 registers are accessed. These registers can, in the most favorable case, be at consecutive addresses and can all be read in a single bus query request, or at addresses 1 to 50, or the registers must be read in separate requests. The contributions of this paper are related to highlighting cases where Modbus modules are incorporated, the addressed registers are not consecutive and the Modbus reads 1 or more holding register records and the ModbusE proposal. As these registers must not be on consecutive addresses, it is necessary to make several requests, defining the function of a Modbus acquisition cycle, adapting the equations to the purpose of the tests carried out, measuring the Modbus communication signals and describing the ModbusE communication concept.
To utilize renewable energies, such as solar energy, distributed power generation systems (DPGS) are widely used, where multiple string inverters are parallelly connected to the grid and often adopt the phase-locked l...
详细信息
The proceedings contain 55 papers. The special focus in this conference is on parallel and distributed Computing: Applications and Technologies. The topics include: A Meta-reinforcement Learning Framework for Ada...
ISBN:
(纸本)9789819642069
The proceedings contain 55 papers. The special focus in this conference is on parallel and distributed Computing: Applications and Technologies. The topics include: A Meta-reinforcement Learning Framework for Adaptive Quadrotor UAV Attitude Control;Securing Energy Transactions for Electric Vehicles: The Blockchain Approach and Encrypted NFTs;optimizing Task Allocation in Heterogeneous Agent Manufacturing systems;MPG: Multi-modal Personal Health Graph for Alzheimer’s Disease Diagnosis;SMAC: A Secure Multi-authority Access Control Scheme with Attribute Unification for Fog Enabled IoT in E-Health;Convolutional Neural Networks Parameter Training for SCM Algorithm Based on Hausdorff Difference;handling Non-stationarity with Distribution Shifts and Data Dependency in time Series Forecasting;the Two-Stage Stochastic Facility Location Game;regularized Non-monotone γ-weakly Submodular Maximization;fed-MoE: Efficient Federated Learning for Mixture-of-Experts Models via Empirical Pruning;WaitIO-Hybrid: Communication for Coupling MPI Programs Among Heterogeneous systems;the Material Delivery Route Prediction Method Based on Deep Reinforcement Learning;privacy-Preserving in Medical Image Analysis: A Review of Methods and Applications;research on Task Migration Problem Based on Link Uncertainty in Adversarial Scenarios;optimizing Production Component Scheduling in Multivariate Industrial Networks with Dynamic Changes in Production Costs;multi-agent Collaboration for time-Sensitive Tasks in Multiple Networked Adversarial Scenarios;containerized Data-Flow Processing for Scalable real-time Analytics on Edge Devices;fast Approximation for Scheduling Malleable Jobs on parallel Batch Machines with Rejection;real-time and In-Situ Temperature Profiling for Determining Detonation of White Dwarf Mergers;accparser: A Standalone OpenACC Parser and Its Usage on Mapping OpenACC to OpenMP Directives;Out-of-Memory GPU Sorting Using Asynchronous CUDA Streams;long-Term and Periodicity-Aware Spatio
This paper introduces XMeta-OS, a meta-operating system specially designed with Linux as its foundation to unify and optimize resource management for the distributed edge-cloud when dynamic use of GPU resources is pos...
详细信息
In the current data-intensive landscape, B+trees are crucial data structures utilized across various fields like databases and web indexing. With the rise of data explosion, the demand for high-performance real-time q...
详细信息
In the current data-intensive landscape, B+trees are crucial data structures utilized across various fields like databases and web indexing. With the rise of data explosion, the demand for high-performance real-time query processing in database systems has surged. For instance, Alibaba's PolarDB and AnalyticDB systems handle massive query volumes and real-time data processing, highlighting the need for efficient solutions. Traditional approaches leveraging GPUs for B+tree performance enhancement have yielded positive results but suffer from high energy consumption, making their widespread deployment in large data centers impractical. This paper introduces GreenB+Tree, an energy-efficient B+tree optimized for the PEZY-SC3s chip, known for its high energy efficiency and MIMD architecture, which mitigates common GPU memory and warp divergence issues. GreenB+Tree innovatively transforms the B+tree structure into two one-dimensional arrays, significantly reducing memory costs and computational overhead. It further incorporates query-agglomerated optimization (QAO) and the persistent data residency strategy (PDRS) to minimize global memory access and enhance cache efficiency. Experimental evaluations demonstrate that GreenB+Tree achieves a throughput of 62.6 Million Queries Per Second per Watt (MQPS/W), outperforming contemporary GPU-based solutions by approximately 4.5 times.
Synthetic Aperture Radar (SAR) tomography is an advanced technique for monitoring deformations of the Earth's surface. However, the computational complexity of SAR tomography algorithms often restricts their appli...
详细信息
The proceedings contains 22 papers from the conference on Proceedings of the 15th workshop on parallel and distributed Simulation. Toipcs discussed include: improving lookahead in parallel discrete event simulations o...
详细信息
The proceedings contains 22 papers from the conference on Proceedings of the 15th workshop on parallel and distributed Simulation. Toipcs discussed include: improving lookahead in parallel discrete event simulations of large-scale applications using compiler analysis;looking ahead of realtime in hybrid component networks;lock-free scheduling of logical processes in parallel simulation;and the dependence list in time wrap.
Online GNN inference has been widely explored by applications such as online recommendation and financial fraud detection systems, where even minor delays can result in significant financial impact. real-time dynamic ...
详细信息
ISBN:
(纸本)9798400714436
Online GNN inference has been widely explored by applications such as online recommendation and financial fraud detection systems, where even minor delays can result in significant financial impact. real-time dynamic graph sampling enables online GNN inference to reflect the latest graph updates in real-world graphs. However, online GNN inference typically demands millisecond-level latency Service Level Objectives (SLOs) as its performance guarantees, which poses great challenges for existing dynamic graph sampling approaches based on graph databases. The issues mainly arise from two aspects: long tail latency due to imbalanced data-dependent sampling and large communication overhead incurred by distributed sampling. To address these issues, we propose Helios, an efficient distributed dynamic graph sampling service to meet the stringent latency SLOs. The key ideas of Helios are 1) pre-sampling the dynamic graph in an event-driven approach, and 2) maintaining a query-aware sample cache to build the complete K-hop sampling results locally for inference requests. Experiments on multiple datasets show that Helios achieves up to 67x higher serving throughput and up to 32x lower P99 query latency compared to baselines.
Practical applications of artificial intelligence increasingly often have to deal with the streaming properties of real data, which, considering the time factor, are subject to phenomena such as periodicity and more o...
详细信息
The proceedings contain 26 papers. The topics discussed include: IGOR: a system for program debugging via reversible execution;static analysis of low-level synchronization;debugging heterogeneous distributedsystems u...
ISBN:
(纸本)0897912969
The proceedings contain 26 papers. The topics discussed include: IGOR: a system for program debugging via reversible execution;static analysis of low-level synchronization;debugging heterogeneous distributedsystems using event-based models of behavior;supporting reverse execution of parallel programs;event-based debugging of object/action programs;integrating tools for debugging and developing multitasking programs;the mach exception handling facility;a real-time monitor for a distributedreal-time operating system;experience with topaz teledebugging;automatic detection of nondeterminacy in parallel programs;a mechanism for efficient debugging of parallel programs;debugging of heterogeneous parallelsystems;initial experiences with a pattern-oriented parallel debugger;high-level debugging in parasight;a program debugger for a systolic array: design and implementation;an integrated approach to parallel program debugging and performance analysis on large-scale multiprocessors;partial orders for parallel debugging;voyeur: graphical views of parallel programs;data path debugging: data-oriented debugging for a concurrent programming language;and concurrent algorithmic debugging.
暂无评论