Asynchronous Execution Research Articles

A multi-level action coupling reinforcement learning approach for online two-stage flexible assembly flow shop scheduling

Multi-product centralized delivery and kitting assembly present significant challenges to hierarchical co-processing in multi-stage manufacturing systems. The combinations of priority dispatching rules at each level are transiently adaptive, and the performance in online scheduling deteriorates rapidly with changing environment. This paper investigates the selection of rule combinations for sustained high-performance responsive scheduling in two-stage flexible assembly flow shop scheduling problem with asynchronous execution and complex decision correlation. A Multi-Level Action Coupling Deep Q-Network (MALC-DQN) approach is proposed for adaptive integrated scheduling in hybrid processing and assembly shops. Firstly, the problem is skillfully established as an event-triggered integrated decision markov decision process. The prioritized batch experience replay mechanism is employed to retain the complete correlation information of key decision sequences. Then, coupling and sequence feature extraction modules are developed to enhance the agent’s ability to perceive execution process and the environment. Furthermore, the multi-level wait-limit mechanism and efficient action filtering mechanism are designed to mitigate ineffective waiting waste and action space explosion during learning. Finally, a series of sophisticated experiments are conducted to validate the effectiveness of the proposed methodology. In 20 actual instances of different sizes, MLAC-DQN outperformed its closest competitor, with a 26.6% improvement in average tardiness. Moreover, extraordinary robustness is demonstrated in 16 sets of experiments involving different configurations of resources, orders, and arrival concentration levels.

Extending Genetic Algorithms with Biological Life-Cycle Dynamics.

In this paper, we aim to enhance genetic algorithms (GAs) by integrating a dynamic model based on biological life cycles. This study addresses the challenge of maintaining diversity and adaptability in GAs by incorporating stages of birth, growth, reproduction, and death into the algorithm's framework. We consider an asynchronous execution of life cycle stages to individuals in the population, ensuring a steady-state evolution that preserves high-quality solutions while maintaining diversity. Experimental results demonstrate that the proposed extension outperforms traditional GAs and is as good or better than other well-known and well established algorithms like PSO and EvoSpace in various benchmark problems, particularly regarding convergence speed and solution qu/ality. The study concludes that incorporating biological life-cycle dynamics into GAs enhances their robustness and efficiency, offering a promising direction for future research in evolutionary computation.

GPUPRSI: GPU implementation of seismic interferometry for retrieving reflection responses from passive source seismic recordings

Passive seismic exploration is an environmentally friendly, economical, and highly accessible exploration method that is widely used in different-scale subsurface imaging. Retrieving reflections from ambient noise is a newly developed passive seismic method that can be used in many fields such as mineral exploration and near-surface imaging, but the interferometry calculation is time-consuming because it requires a longer period of data acquisition to improve the signal-to-noise ratio of the retrieved reflections. In this study, we introduced a graphical processing unit (GPU)- based implementation of seismic interferometry for retrieving reflection responses from passive source seismic recordings. Because all traces are involved in the computation process, but the size of passive source data often exceeds available memory, repeated disk reads lead to a decrease in computational efficiency. We design a strategy of grouping computations followed by stacking to minimize disk input and output, simultaneously keeping the memory requirements low. Passive source data is read and written only once, and there is no requirement for the memory size to be greater than the data size. Additionally, acceleration technologies such as asynchronous execution, asynchronous memory transfer, and GPU-accelerated libraries are used. We test the efficiency using short data where the number of sampling points per trace is on the order of 30,000, and long data, where the number of sampling points per trace is on the order of 30,000,000. For short data, the average speedup is 4 compared with a multi-core central processing unit (CPU); for long data, the speedup can reach 24. The GPU-based implementation of interferometry greatly reduces the calculation time for retrieving reflections from passive source seismic recordings, providing a solution to the problem of large calculation in three-dimensional (3D) passive source reflection exploration.

Embedded model control of networked control systems: An experimental robotic application

In Networked Control System (NCS), the absence of physical communication links in the loop leads to relevant issues, such as measurement delays and asynchronous execution of the control commands. In general, these issues may significantly compromise the performance of the NCS, possibly causing unstable behaviours. This paper presents an original approach to the design of a complete digital control unit for a system characterized by a varying sampling time and asynchronous command execution. The approach is based on the Embedded Model Control (EMC) methodology, whose key feature is the estimation of the disturbances, errors and nonlinearities affecting the plant to control and their online cancellation. In this way, measurement delays and execution asynchronicity are treated as errors and rejected up to a given frequency by the EMC unit. The effectiveness of the proposed approach is demonstrated in a real-world case-study, where the NCS consists of a differential-drive mobile robot (the plant) and a control unit, and the two subsystems communicate through the web without physical connection links. After a preliminary verification using a high-fidelity numerical simulator, the designed controller is validated in several experimental tests, carried out on a real-time embedded system incorporated in the robotic platform.

Sensitivity-based distributed model predictive control: synchronous and asynchronous execution compared to ADMM

Abstract This paper presents synchronous and asynchronous formulations of a sensitivity-based algorithm for solving distributed continuous-time nonlinear optimal control problems with a neighbor-affine structure. Sensitivities are defined in an optimal control context and utilized to ensure coordination between agents. By using delayed data in the asynchronous formulation, idle times of agents are reduced and faster execution of the algorithm is achieved. The convergence behavior w.r.t. topology and coupling strength is investigated in a numerical simulation and the execution time is evaluated on distributed hardware with communication over TCP in comparison to the alternating direction method of multipliers (ADMM) algorithm.

Open-source complex-geometry 3D fluid dynamics for applications with unpredictable heterogeneous dynamic high-performance-computing loads

We discuss the software implementation of a finite element method, intended for the simulation of complex-geometry 3D flows, using hardware resources effectively, including problems with a priori unknown, heterogeneous, and dynamic parallel computational load. Our fundamental choices of data structures, algorithm, and software design specifically target engineering resolution and accuracy requirements. Some of these choices are: unstructured grids (to explicitly resolve complex 3D geometries), tetrahedra-only computational elements (to enable automatic mesh generation), edge-based finite element scheme (to reduce indirect addressing for increased performance), distributed-memory parallel computing paradigm (to enable large problems), and Charm++ (https://charmplusplus.org) as the runtime system (to effectively use computing resources even in the presence of hardware heterogeneities and dynamic application requirements). We discuss aspects of the implementation that enable exercising unique features of the runtime system, e.g., the single Charm++ programming abstraction, overdecomposition, asynchronous execution, latency-hiding parallel communication and computation, task-parallelism, and dynamic load balancing via object migratability. Multiple test problems are used to verify and validate the numerical solutions and computational performance and scalability to high-performance computing environments are discussed. For maximum transparency and reproducibility, and to encourage future research, development, and use, the full source code, together with regression tests and documentation, is publicly available at https://xyst.cc.

Drug co‐administration in anesthesia using event‐based MPC

AbstractIn this article, a predictive event‐based control architecture for multivariable drug infusion in the anesthesia process is presented. The control scheme considers a multiple‐input single‐output process, where the depth of hypnosis is the controlled variable and the infusion rates of propofol and remifentanil, which are co‐administered by imposing a fixed ratio between them, are the control variables. The control system is built on the top of a nonlinear pharmacokinetic/pharmacodynamic model for drug co‐administration that reflects the super‐additive effect of both drugs on the bispectral index scale (BIS), which is a measure of the hypnotic state of the patient. Further, in order to compensate the nonlinear behavior of the system, it exploits an external predictor that is designed to increase the robustness to inter‐/intra‐patient variability. The overall controller parameters are tuned by applying an optimization method on a representative set of virtual patients. The main feature of the proposed control algorithm consists in its asynchronous execution, which can be beneficial in reducing control signal changes due to the noisy measurements of the BIS. The evaluation of the analyzed control architecture through a simulation study reveals that control action variations can be reduced by about 48% on average with respect to its time‐based counterpart. This reduction is obtained at the expense of a lower control accuracy, resulting in a degradation of 18% in terms of integrated absolute error. Results are acceptable from a clinical practice perspective proving that the proposed approach is a feasible alternative to classical time‐based predictive controllers.

Effect of asynchronous execution and imperfect communication on max-sum belief propagation

Effect of asynchronous execution and imperfect communication on max-sum belief propagation

DNA-based programmable gate arrays for general-purpose DNA computing.

The past decades have witnessed the evolution of electronic and photonic integrated circuits, from application specific to programmable1,2. Although liquid-phase DNA circuitry holds the potential for massive parallelism in the encoding and execution of algorithms3,4, the development of general-purpose DNA integrated circuits (DICs) has yet to be explored. Here we demonstrate a DIC system by integration of multilayer DNA-based programmable gate arrays (DPGAs). We find that the use of generic single-stranded oligonucleotides as a uniform transmission signal can reliably integrate large-scale DICs with minimal leakage and high fidelity for general-purpose computing. Reconfiguration of a single DPGA with 24 addressable dual-rail gates can be programmed with wiring instructions to implement over 100 billion distinct circuits. Furthermore, to control the intrinsically random collision of molecules, we designed DNA origami registers to provide the directionality for asynchronous execution of cascaded DPGAs. We exemplify this by a quadratic equation-solving DIC assembled with three layers of cascade DPGAs comprising 30 logic gates with around 500 DNA strands. We further show that integration of a DPGA with an analog-to-digital converter can classify disease-related microRNAs. The ability to integrate large-scale DPGA networks without apparent signal attenuation marks a key step towards general-purpose DNA computing.

Edge‐Friend: Fast and Deterministic Catmull‐Clark Subdivision Surfaces

AbstractWe present edge‐friend, a data structure for quad meshes with access to neighborhood information required for Catmull‐Clark subdivision surface refinement. Edge‐friend enables efficient real‐time subdivision surface rendering. In particular, the resulting algorithm is deterministic, does not require hardware support for atomic floating‐point arithmetic, and is optimized for efficient rendering on GPUs. Edge‐friend exploits that after one subdivision step, two edges can be uniquely and implicitly assigned to each quad. Additionally, edge‐friend is a compact data structure, adding little overhead. Our algorithm is simple to implement in a single compute shader kernel, and requires minimal synchronization which makes it particularly suited for asynchronous execution. We easily extend our kernel to support relevant Catmull‐Clark subdivision surface features, including semi‐smooth creases, boundaries, animation and attribute interpolation. In case of topology changes, our data structure requires little preprocessing, making it amendable for a variety of applications, including real‐time editing and animations. Our method can process and render billions of triangles per second on modern GPUs. For a sample mesh, our algorithm generates and renders 2.9 million triangles in 0.58ms on an AMD Radeon RX 7900 XTX GPU.

The Story of GraphLab - From Scaling Machine Learning to Shaping Graph Systems Research (VLDB 2023 Test-of-Time Award Talk)

The GraphLab project spanned almost a decade and had profound academic and industrial impact on large-scale machine learning and graph processing systems. There were numerous papers written describing the innovations in GraphLab including the original vertex-centric [8] and edge-centric [3] programming abstractions, high-performance asynchronous execution engines [9], out-of-core graph computation [6], tabular graph-systems [4], and even new statistical inference algorithms [2] enabled by the GraphLab project. This work became the basis of multiple PhD theses [1, 5, 7]. The GraphLab open-source project had broad academic and industrial adoption and ultimately lead to the launch of Turi. In this talk, we tell the story of GraphLab, how it began and the key ideas behind it. We will focus on the approach to achieving scalable asynchronous systems in machine learning. During our talk, we will explore the impact that GraphLab has had on the development of graph processing systems, graph databases, and AI/ML; Additionally, we will share our insights and opinions into where we see the future of these fields heading. In the process, we highlight some of the lessons we learned and provide guidance for future students.

Information systems for automating work flows: a comparative analysis

Workflow is a sequence of repeated and controlled steps aimed at completing a specific task or work. The urgency of the task of optimizing such works contributed to the development of methods and means of operations research to optimize such processes for the needs of various subject areas. Information technologies to support such workflow are workflow engines that enable faster automation, ensure compliance with norms and standards, formalize business processes, improve communication, etc. The workflow management systems are divided into the following categories: automation using robotics, intelligent integration platforms, intelligent business process management systems, open-source engines, the cloud-supported, as well as data flow processing systems. For a more detailed analysis, it is advisable to choose the following engines and services: built-in, cloud-oriented, and those to support both scenarios (jBPM, Camunda, Zeebe, Amazon Step Functions). It is appropriate to define evaluation criteria and compare such workflow automation systems to form further recommendations regarding their selection and application. Such criteria are development activity, stability and history of commercial use, versioning support, standards support, support for timers and asynchronous execution, support for human-oriented and manual tasks, integration with other solutions, monitoring and logging, scaling, cloud support, the possibility of scanning in private infrastructure, the presence of a visual interface, the convenience of local development and testing, open source code, the necessary programming for implementation and cost. A comparison of advantages and disadvantages can be used to decide on a workflow automation system.

Making applications faster by asynchronous execution: Slowing down processes or relaxing MPI collectives

Comprehending the performance bottlenecks at the core of the intricate hardware–software interactions exhibited by highly parallel programs on HPC clusters is crucial. This paper sheds light on the issue of automatically asynchronous MPI communication in memory-bound parallel programs on multicore clusters and how it can be facilitated. For instance, slowing down MPI processes by deliberate injection of delays can improve performance if certain conditions are met. This leads to the counter-intuitive conclusion that noise, independent of its source, is not always detrimental but can be leveraged for performance improvements. We employ phase-space graphs as a new tool to visualize parallel program dynamics. They are useful in spotting certain patterns in parallel execution that will easily go unnoticed with traditional tracing tools. We investigate five different microbenchmarks and applications on different supercomputer platforms: an MPI-augmented STREAM Triad, two implementations of Lattice-Boltzmann fluid solvers (D3Q19 and SPEChpc D2Q37), the LULESH and HPCG proxy applications.

GeoGauss: Strongly Consistent and Light-Coordinated OLTP for Geo-Replicated SQL Database

Multinational enterprises conduct global business that has a demand for geo-distributed transactional databases. Existing state-of-the-art databases adopt a sharded master-follower replication architecture. However, the single-master serving mode incurs massive cross-region writes from clients, and the sharded architecture requires multiple round-trip acknowledgments (e.g., 2PC) to ensure atomicity for cross-shard transactions. These limitations drive us to seek yet another design choice. In this paper, we propose a strongly consistent OLTP database GeoGauss with full replica multi-master architecture. To efficiently merge the updates from different master nodes, we propose a multi-master OCC that unifies data replication and concurrent transaction processing. By leveraging an epoch-based delta state merge rule and the optimistic asynchronous execution, GeoGauss ensures strong consistency with light-coordinated protocol and allows more concurrency with weak isolation, which are sufficient to meet our needs. Our geo-distributed experimental results show that GeoGauss achieves 7.06X higher throughput and 17.41X lower latency than the state-of-the-art geo-distributed database CockroachDB on the TPC-C benchmark.

Incremental Connected Component Detection for Graph Streams on GPU

Studies on the real-time detection of connected components in graph streams have been carried out. The existing connected component detection method cannot process connected components incrementally, and the performance deteriorates due to frequent data transmission when GPU is used. In this paper, we propose a new incremental processing method to solve the problems found in the existing methods for detecting connected components on GPUs. The proposed method minimizes the amount of data to be sent to the GPU by determining the subgraph affected by the graph stream update and by detecting the part to be recalculated. We consider the number of vertices to quickly determine the connected components of a graph stream on the GPU. An asynchronous execution method is used to shorten the transfer time between the CPU and the GPU according to real-time graph stream changes. In order to show that the proposed method provides fast incremental connected component detection on the GPU, we evaluated its performance using various datasets.

Machine learning based asynchronous computational framework for generalized Kalman filter

SummaryAlthough the Kalman filter algorithms are well suited to be executed on most digital systems, they become slow when applied to large‐scale dynamic systems. Therefore, efficient execution of Kalman filter for the time‐critical and large‐scale applications is of the essence. This work aims to address this necessity by developing a novel framework to improve the performance of a generalized Kalman filter with unknown inputs (GKF‐UI) using multithreaded‐multicore processors and machine learning (ML) classification methods. An asynchronous execution model based on OpenMP message‐passing framework is developed and integrated with a novel supervised ML‐based thread classifier for the GKF‐UI algorithm to enhance execution efficiency. The experimental results show that the proposed approach can achieve up to 35.5× speedup over the serial single‐threaded implementations with no losses in the accuracy or changes to the generality of the filter structure. Moreover, this framework can play a significant role in realizations of computational advantages in large‐scale systems as well as for the time‐critical prediction applications.

Secure HPC: A workflow providing a secure partition on an HPC system

Driven by the progress of data and compute-intensive methods in various scientific domains, there is an increasing demand from researchers working with highly sensitive data to have access to the necessary computational resources to be able to adapt those methods in their respective fields. To satisfy the computing needs of those researchers cost-effectively, it is an open quest to integrate reliable security measures on existing High Performance Computing (HPC) clusters. The fundamental problem with securely working with sensitive data is, that HPC systems are shared systems that are typically trimmed for the highest performance—not for high security. For instance, there are commonly no additional virtualization techniques employed, thus, users typically have access to the host operating system. Since new vulnerabilities are being continuously discovered, solely relying on the traditional Unix permissions is not secure enough. In this paper, we discuss Secure HPC, a workflow allowing users to transfer, store and analyze data with the highest privacy requirements. Our contributions are the design of a multi-node secure workflow with parallel I/O, a strict security model enforced by the system and network features, and lastly the demonstration of a medical use case. In our experiments, we see an advantage in the asynchronous execution of IO requests in dm_crypt, while reaching 80% of the ideal performance. When comparing eCryptFS with GoCryptFS as two representative filesystem-level encryption stacks, eCryptFS was twice as fast. In a real use case, we observed on average 97% of the native performance.

SyncNN: Evaluating and Accelerating Spiking Neural Networks on FPGAs

Compared to conventional artificial neural networks, spiking neural networks (SNNs) are more biologically plausible and require less computation due to their event-driven nature of spiking neurons. However, the default asynchronous execution of SNNs also poses great challenges to accelerate their performance on FPGAs. In this work, we present a novel synchronous approach for rate-encoding-based SNNs, which is more hardware friendly than conventional asynchronous approaches. We first quantitatively evaluate and mathematically prove that the proposed synchronous approach and asynchronous implementation alternatives of rate-encoding-based SNNs are similar in terms of inference accuracy, and we highlight the computational performance advantage of using SyncNN over an asynchronous approach. We also design and implement the SyncNN framework to accelerate SNNs on Xilinx ARM-FPGA SoCs in a synchronous fashion. To improve the computation and memory access efficiency, we first quantize the network weights to 16-bit, 8-bit, and 4-bit fixed-point values with the SNN-friendly quantization techniques. Moreover, we encode only the activated neurons by recording their positions and corresponding number of spikes to fully utilize the event-driven characteristics of SNNs, instead of using the common binary encoding (i.e., 1 for a spike and 0 for no spike). For the encoded neurons that have dynamic and irregular access patterns, we design parameterized compute engines to accelerate their performance on the FPGA, where we explore various parallelization strategies and memory access optimizations. Our experimental results on multiple Xilinx ARM-FPGA SoC boards demonstrate that our SyncNN is scalable to run multiple networks, such as LeNet, Network in Network, and VGG, on various datasets such as MNIST, SVHN, and CIFAR-10. SyncNN not only achieves competitive accuracy (99.6%) but also achieves state-of-the-art performance (13,086 frames per second) for the MNIST dataset. Finally, we compare the performance of SyncNN with conventional CNNs using the Vitis AI and find that SyncNN can achieve similar accuracy and better performance compared to Vitis AI for image classification using small networks.

Project management of «unicorn» companies under digitalization conditions

Methods. The results were obtained through the use of the following methods: comparison – when determining the advantages of different approaches to project management in a specific historical context; statistical analysis – when determining the rates of digitization of business activity in Ukraine and in the world; induction and deduction – when forming recommendations regarding the construction of the project management system of high-tech companies. Results. Trends in the change of project management model in startup organizations have been identified. They are associated with their transition to agile management methods, in particular, those modern project management models that have been developed in the leading countries of the world. Various project management methodologies used by «unicorn» companies are analyzed. It has been established that the desire to combine the interests of various groups of interested parties in the conditions of defined time, financial and technological possibilities leads to the transition from traditional cascade methods of planning and organizing activities to the division of works into separate components with their asynchronous execution. Novelty. The research has demonstrated an increase in the number of business structures that are digitizing their own business model, which is accompanied by a gradual rejection of classic administrative approaches (in particular, those that were inherent in cascade planning), as well as a transition to flexible forms of project management, known as agile. The key drivers of digital transformation of companies are described. Practical value. There is offered the description of the trends in the digital transformation of business structures by adapting startup companies of their own business models to dynamic changes in the market environment in both the global and local dimensions. The application of flexible management methods allows high-tech companies to launch new projects, using outsourcing and applying an incremental approach to the development of new products. Thanks to this, companies introduce «beta versions» of products much earlier and with less expenditure of resources, gradually improving them.

Scalable Artificial Intelligence for Earth Observation Data Using Hopsworks

This paper introduces the Hopsworks platform to the entire Earth Observation (EO) data community and the Copernicus programme. Hopsworks is a scalable data-intensive open-source Artificial Intelligence (AI) platform that was jointly developed by Logical Clocks and the KTH Royal Institute of Technology for building end-to-end Machine Learning (ML)/Deep Learning (DL) pipelines for EO data. It provides the full stack of services needed to manage the entire life cycle of data in ML. In particular, Hopsworks supports the development of horizontally scalable DL applications in notebooks and the operation of workflows to support those applications, including parallel data processing, model training, and model deployment at scale. To the best of our knowledge, this is the first work that demonstrates the services and features of the Hopsworks platform, which provide users with the means to build scalable end-to-end ML/DL pipelines for EO data, as well as support for the discovery and search for EO metadata. This paper serves as a demonstration and walkthrough of the stages of building a production-level model that includes data ingestion, data preparation, feature extraction, model training, model serving, and monitoring. To this end, we provide a practical example that demonstrates the aforementioned stages with real-world EO data and includes source code that implements the functionality of the platform. We also perform an experimental evaluation of two frameworks built on top of Hopsworks, namely Maggy and AutoAblation. We show that using Maggy for hyperparameter tuning results in roughly half the wall-clock time required to execute the same number of hyperparameter tuning trials using Spark while providing linear scalability as more workers are added. Furthermore, we demonstrate how AutoAblation facilitates the definition of ablation studies and enables the asynchronous parallel execution of ablation trials.