KoHPCG is a performance-portable HPCG benchmark program developed using the Kokkos programming model. The Reference HPCG benchmark is constrained by memory-bound kernels that restrict performance across various architectures. This paper details a thorough implementation of core HPCG kernels, including Dot Product (DDOT), WAXPBY, Sparse Matrix Vector Multiplication (SpMV), Symmetric Gauss-Seidel (SymGS), and Multigrid (MG), utilizing the Kokkos::Views, parallel constructs, execution and memory space abstraction. Evaluated on Intel Xeon Phi (KNL) and Xeon Skylake (SKL) processors with a maximum of 16 nodes, KoHPCG achieves significant performance improvement up to $11.7 \times$ acceleration in MG on SKL, alongside a $17.3 \times$ MG improvement on KNL. The overall HPCG performance increases by as much as $3.2 \times$ on SKL and attains $4.1 \times$ on KNL compared to the Reference HPCG implementation, on problem sizes of $192^{3}$ and $160^{3}$ on SKL and KNL, respectively. Furthermore, on a larger problem size of $320^{3}$, KoHPCG attains an overall HPCG performance improvement of $5.1 \times$ on SKL. In addition to the overall performance improvement in HPCG, we reported the kernel-level performance that exposes the little and no performance improvement in SpMV and WAXPBY respectively, and increased memory consumption by using the Kokkos data structure, thereby highlighting further optimisation opportunities for future work.

Graphic Processing Units (GPUs) are known for the large computing capabilities they offer users compared to traditional CPUs. However, the issue of resource under-utilization is becoming more apparent as more and more applications are unable to saturate modern GPUs which have even higher processing capabilities. While concurrency mechanisms like hardware partitioning have resulted in better utilization compared to deployments without sharing, the issue of resource under-utilization still persists even in deployment scenarios where applications are executed on the smallest GPU partitions of modern GPUs. Software partitioning on the other hand, does not guarantee isolation during executions leading to issues of interference and consequently limiting the number of applications which can be run concurrently. Leveraging both software and hardware resource partitioning schemes in an effort to mitigate resource under-utilization issues is yet to be fully explored. In this paper, we evaluate the predictions of a proposed linear regression model relative to actual executions. The results of our experiments show that whilst our approach accurately estimates performance for sharing differently-sized GPU partitions among diverse applications based on each application’s characteristics, it also improves utilization and reduces resource wastage.

Cache management is a significant aspect of executing applications on GPUs. With the advancements in GPU architecture, issues such as data reuse, cache line eviction and data residency are to be considered for optimal performance. Frequency of data access from global memory has significant impacts on the performance of the application with increased latencies. However, the L2 cache data residency feature by NVIDIA promises to reduce the overheads associated with frequent data accesses. Through the information extracted from static profiling analysis, we quantitatively analyzed the frequency of data reuse by threads to determine whether an application has frequent data accesses or not. We also estimated the size of access policy window from which persistent data should be cached to avoid stalling of warps. Also with our proposed approach, we observed that L1 cache load throughput increased by 2.75% for GEMM, 0.33% for 2DConv St and 0.46% for 2DConv Large respectively as data was resident in the L2 cache.

The most scientific and numerical problems can be solved using the system of equations in linear algebra. Matrix-matrix multiplication is the foundation of linear algebra equations, and its optimization has an impact on the overall performance of a system. ScaLAPACK has established itself as the industry standard for dense linear algebraic computations, developed 30 years ago. Owing to advancements in microprocessor architectures, it is difficult to fully utilize the hardware capabilities of legacy software systems on modern architectures and achieve the maximum performance. In this study, we analyzed the effects of matrix size, register blocking parameters, and thread distribution on the performance, and improved our previously implemented matrix-matrix multiplication routine for matrix-panel multiplication, which performed well for large-sized square matrices. We also presented the ScaLAPACK QR factorization performance by replacing the double-precision general matrix-matrix multiplication routine (DGEMM) of ScaLAPACK with our matrix-matrix multiplication routine for a single node Intel Xeon Phi Knights Landing processor.

Summary Edge computing focuses on processing near the source of the data. Edge computing devices using the Tegra SoC architecture provide a physically distinct GPU memory architecture. In order to take advantage of this architecture, different modes of memory allocation need to be considered. Different GPU memory allocation techniques yield different results in memory usage and execution times of identical applications on Tegra devices. In this article, we implement several GPU application benchmarks, including our custom CFD code with unified, pinned, and normal host/device memory allocation modes. We evaluate and compare the memory usage and execution time of such workloads on edge computing Tegra system‐on‐chips (SoC) equipped with integrated GPUs using a shared memory architecture, and non‐SoC machines with discrete GPUs equipped with distinct VRAM. We discover that utilizing normal memory allocation methods on SoCs actually use double the required memory because of unnecessary device memory copies, despite being physically shared with host memory. We show that GPU application memory usage can be reduced up to 50%, and that even performance improvements can occur just by replacing normal memory allocation and memory copy methods with managed unified memory or pinned memory allocation.

The advent of GPGPU (General-Purpose Graphic Processing Unit) containers enlarges opportunities of acceleration and easy-to-use in clouds. However, there is still lack of research on utilizing efficiently GPU resource and managing multiple applications at the same time. Co-execution of applications without understanding applications' execution characteristics may result in low performance caused by their interference problems. To solve the problem, this paper defines resource metrics that causes performance degradation when sharing resource. We calculate the degree of interference during concurrent execution of multi applications using a ML (Machine Learning) method with the metrics. The experiments show that the execution of interference aware groups improves 7% in execution time compared to non-interference aware group in overall. For a workload consisting of several applications, the overall performance was improved by 18% and 25%, respectively, when compared to SJF and random.

Recently, efficient execution of applications on Graphic Processing Unit(GPU) has emerged as a research topic to increase overall system throughput in cluster environment. As a current cluster orchestration platform using GPUs only supports an exclusive execution of an application on a GPU, the platform may not utilize resource of GPUs fully relying on application characteristics. Nonetheless, co-execution of GPU applications leads to interference coming from resource contention among applications. If diverse resource usage characteristics of GPU applications are not deliberated, unbalanced usage of computing resources and performance degradation could be induced in a GPU cluster. This study introduces Co-scheML for co-execution of various GPU applications such as High Performance Computing (HPC), Deep Learning (DL) Training, and DL Inference. Interference model is constructed by applying Machine Learning (ML) model with GPU metrics since predicting interference has a difficulty. Predicted interference is utilized and deployment of an application is determined by Co-scheML scheduler. Experimental results of the Co-ScheML strategy show that average job completion time is improved by 23%, and the makespan is shortened by 22% in average, as compared to baseline schedulers.

Graphics Processing Units (GPUs) are getting popularly utilized for multi-purpose applications in order to enhance highly performed parallelism of computation. As memory virtualization methods in GPU nodes are not efficiently provided to deal with diverse memory usage patterns for these applications, the success of their execution depends on exclusive and limited use of physical memory in GPU environments. Therefore, it is important to predict a pattern change of GPU memory usage during runtime execution of an application. Data provenance extracted from application characteristics, GPU runtime environments, input, and execution patterns from runtime monitoring, is defined for supporting application management to set runtime configuration and predict an experimental result, and utilize resource with co-located applications. In this paper, we define data provenance of an application on GPUs and manage data by profiling the execution of CUDA scientific applications. Data provenance management helps to predict execution patterns of other similar experiments and plan efficient resource configuration.

Data of scientific workflow applications are tend to be distributed over many data centers to be effectively stored, retrieved, and transferred among them. The result of an experiment with those data shows diverse execution performance depending on the placement of input and intermediate data which are generated during application execution. However, initial data placement strategy would not be the best plan for long running experiments because of the dynamic change of resource condition time to time. We propose an adaptive data placement strategy considering dynamic resource change for efficient data-intensive applications. The strategy consists of two stages that group the datasets in data centers during the build- time stage and dynamically clusters every time newly generated datasets repeatedly to the most appropriate data centers during runtime stage, which is based on task dependency, intense degree of data usage, and just-in-time resource availability. Just-in-time data placement coming with task execution is more efficient than the one with initialization stage of experiments in the aspect of resource utilization. Experiments show that data movement can be effectively reduced while the workflow is running.

Supporting end-to-end network slices and services across operators has become an important use case of study for 5G networks as can be seen by 5G use cases published in 3GPP, ETSI as well as NGMN. This paper presents the in- depth architecture, implementation and experiment on a multi-domain orchestration framework that is ab le to deploy such multi-operator service as well as monitor the service for SLA compliance. Our implemented architecture allows operators to abstract their sensitive details while exposing the relevant amount of information to support inter-operator slice creation. Our experiment shows that the implemented framework is capable of creating services across operators while fulfilling the respective service requirements.

Cloud computing platform is becoming one of the most desired environments for the emerging applications as well as traditional application domains. In this regards, there have been lots of efforts to develop and redesign the framework to deliver solutions that meet a variety of requirements including resilience, performance, and agility. This paper presents a design and implementation of iSDF, an integrated software-defined computing framework for scientific experiments. The iSDF aims at reinforcing traditional cloud-based scheme by supporting application-aware scheduling that avoids resource starvation and contentions for software-defined computing. To demonstrate the effectiveness of the iSDF, we evaluate it by performing several experiments based on three scenarios. The results show that our framework can achieve a 51% increase on resource utilization and average makespan time a 19% improvement compared to the conditions that do not include the policy. Thus, we proved that the framework can contribute to increase utilization for shared resources and to optimize overall makespan time, eventually affecting overall throughput of jobs.

As the development of Internet and mobile device increases, there is a correspondingly increasing amount of data produced by users of such technology worldwide. It is thus essential to analyze such massive amounts of data reflective of the big data era. Recently, Apache Spark has become popular for analyzing big data, since it can process streaming data and support real-time in-memory computing. Also, it is known to execute applications faster than traditionally used Hadoop. Also cloud technology provides flexible resource utilization environment on-demand. When analyzing big data using Spark in existing environments, it is difficult to provision resources according to the system's changing environment and the influence of other users' executions. Using cloud technology however, it is possible to provision resources more effectively for the execution of jobs through dynamic resource provision methods. In this paper, we propose an auto-scaling framework with corresponding algorithms to manage resources dynamically in virtual environments, in order to meet user-specified SLA (Service Level Agreement) given a set of limited resources. Our experimental results on Spark in OpenStack demonstrate the effectiveness of scaling resources to satisfy user SLAs.

High Performance Computing (HPC) applications require systems with environments for maximum use of limited resources to facilitate efficient computations. However, these systems are faced with a large trade-off between efficient resource allocation and minimum execution times for the applications executed on them. Also, deploying applications in newer environments is exacting. To alleviate this challenge, container-based systems are recently being deployed to reduce the trade-off. In this paper, we investigate container-based technology as an efficient virtualization technology for running high performance scientific applications. We select Docker as the container-based technology for our test bed. We execute autodock3, a molecular modeling simulation software mostly used for Protein-ligand docking, in Docker containers and VMs created using OpenStack. We compare the execution times of the docking process in both Docker containers and in VMs.

Scientists gain benefits from scalable resource provisioning on-demand, and various computing environments by using cloud computing resources for their applications. However, many cloud computing providers offer their cloud resources according to their own rules. The descriptions of various cloud resources also differ for each vendor. Subsequently, it becomes difficult to find suitable cloud resources given the characteristics of an application. Scientists therefore, select resources used in previous experiments or best performance resources without considering the characteristics of their applications. There is the need to standardize notations to support simple selection of cloud resources without the constraints of providers. Intercloud can use cloud resources without considering cloud providers in hybrid cloud environments. Intercloud project has been studied for interoperability and creates mOSAIC ontology to conceptualize various resources. However, the mOSAIC ontology attributes are limited when considering characteristics of an application. We propose a semantic engine to provide semantic cloud resource services in intercloud environment. We define a rule of categorized resource description with reference to mOSAIC ontology attributes which includes added factors needed to represent application characteristics. We also develop a semantic engine which can choose semantically similar cloud resources using statistical analysis while considering the characteristics of an application. The semantic engine can also classify resources dynamically according to application specifications.

Scientists can ease to conduct large-scale scientific computational experiments over cloud environment according to an appearance of Science Clouds. Cloud computing enables applications to apply on-demand and scalable resources dynamically. It is necessary for Many Task Computing (MTC) to offer high performance resources in a long phase and certificate stable executions of applications even dramatic changes of vital status of physical resources. Auto-scaling on virtual machines provides integrated and efficient utilization of cloud resources. VM Auto-scaling schemes have been actively studied as effective resource management in order to utilize large-scale data center in a good shape. However, most of the existing auto-scaling methods just simply support CPU utilization and data transfer latency. It is needed to consider execution deadline or characteristics of an application. We propose an auto-scaling method, guaranteeing the execution of an application within deadline. It can handle two types of job patterns; Bag-of-Tasks jobs or workflow jobs. We simulate a variable index computation application in hybrid cloud environment. The results of the simulation show the method can dynamically allocate resources considering deadline.

As computing technologies develop, various studies for e-Science have been actively developed for many years. The advent of Cloud computing, for example, enables scientists to expand their research environments over supercomputers to on-demand and scalable resources. However, several performance drawbacks on the cloud computing can cause considerable obstacles in scientific domains, despite its strong merits. A hybrid infrastructure which consists of the existing and cloud architecture thus can produce a synergy effect for utilizing resources efficiently. In this paper, we proposed hybrid infrastructure-supported a scientific workflow environment for Aerodynamics design and demonstrated its superiority. Hybrid infrastructure covers grid and private cloud computing in this paper. Especially, we focused on improving performance by supporting hybrid infrastructure and efficient usages of physical resources. Since it offers diverse types of computing infrastructures including cloud computing, it can serve elastic resources regardless of the number of tasks for experiments or limitations of space and can reduce costs of time as well as budget during simulations.

Due to diverse characteristics of application domains, most experts dealing with scientific computational simulations require a specialized and customized Problem Solving Environment (PSE) with easy-to-use graphical user interfaces integrated with large-scale computing resources for their experiments. It is also important to provide adequate middleware support to drag out best performance of an application from the environment. To fulfill these requirements, we propose a workflow-enabled integrated e-Science portal system supporting interactivities for aerodynamic computing simulations, which are driven to diverse multiple execution stages depending on convergence status of variables. The portal allows scientists to manage whole life-cycle of experiments from designing, deploying, executing, and monitoring a scientific application in aerodynamic research on e-Science environment. Scientists easily extend various types of their experiments with less risk of design errors and save their time for learning curve to handle computing resources to optimize the performance of experiments. Workflow editor presents interactive, high-level web-based front end to scientist and monitoring service shows the step of the execution of workflow in real-time so that scientists in any level of expertise can check the intermediate data during experiments.

The advent of cloud computing makes scientists to extend their research environments over supercomputers to on-demand and dynamically scalable resources. Science cloud becomes a trend in various scientific domains these days. However, it is difficult to provide optimal job execution environment rapidly and dynamically depending on user's demands. Therefore, it is very important to predict user's requirements and to prepare execution environment in advance. In addition, it needs scheduling mechanisms for virtual machines to provide some level of guaranteed performance of a user application. In this paper, we propose a cloud resource provisioning model using statistical analysis of job history. In this model, we use job history which is generated from many application executions and identifies characteristics of an application by applying statistical analysis. We utilize a statistical technique, PCA (Principal Component Analysis), to analyze execution history of applications and to extract the factors which contribute much to execution time. The effective factors are used for selecting reference job profile and then VM is deployed on the selected node based on the reference profile. An application is executed on chosen nodes and its performance result is incorporated into job history with the purpose of evaluating profile's credit. As a result, this model can provide efficient management of cloud resource for a service provider and reduce management overhead on cloud.

Grids have ensured to solve a large-scale complex scientific problem with utilizing many heterogeneous resources that are geographically separated. However, as grid resources and scientific applications become diverse, the difficulty of their execution management and control has been increased. In this paper, we propose JARS (Job history-based Adaptive grid Resource Selection) system and evaluate with different scenarios. Specifically, statistical technique, Plackett-Burman design with fold-over, is applied on analyzing grid environments and job history. Through JARS system, accumulated knowledge of applications' execution can be utilized in practical resource selection. As a proof-of-concept, we analyzed job history from an aerospace research grid system and obtained characteristics of grid resources and applications. The simulations present that resources with acceptable performance and reliability are selected under heavy resource failures.

Cyberenvironment has emerged as a next generation cyberinfrastructure to support 21st century scientific research and education. In this paper, we introduce an on-going cyberenvironment project named e-AIRS, anabbreviation of 'e-science aerospace integrated research system'. e-AIRS is a cyberinfrastructure-based portal system, and integrates a set of tools and services to support the aerodynamics engineering processes as a first step toward building a cyberenvironment for engineering cyber education. The system helps students to understand the full simulation process of aerodynamics engineering. Moreover, the system can be widely applied to other engineering fields too, such as ships, civil engineering, and etc. We conducted a survey of about 230 students about 5 topics at 7 classes of five universities. According to the survey, 94 percent said the system helped to understand the whole process of CFD, and more than 80 percent of the students said e-AIRS environment provided a good functionality, convenience, and user interface.

Management of large-scale network-centric systems (NCS) and their applications is an extremely complex and challenging task due to factors such as centralized management architectures, lack of coordination and compatibility among heterogeneous network management systems, and the dynamic characteristics of networks and application bandwidth requirements, just to name a few. The goal of this research is to develop a hierarchical framework to achieve end-to-end intelligent proactive network management system that can be used to manage large-scale network-centric systems and their applications. This framework provides the ability to write management programs to manage any required function or property (performance, high assurance, fault, quality of service, etc.) of the network-centric systems and their applications during all the phases of their operations. Our ultimate objective is to consider the management of network-centric systems and applications starting from the design phase and forward rather than being afterthought process. In this paper, we present a framework to develop proactive and adaptive management services and an implementation of a proactive application management system (PAMS) based on that framework. Our implementation approach utilizes delegated mobile agents to implement the management functions required by any network-centric system and/or application. We also present experimental results and evaluation of the management services offered by the PAMS prototype.

Current advances in networking, computing, software and web technologies have led to an explosive growth in the development of networked applications. Management of large-scale networks and their applications is an extremely complex task due to factors such as centralized management architectures, lack of coordination and compatibility among heterogeneous network management systems, and dynamic characteristics of networks and application bandwidth requirements. We do need to develop an integrated network management paradigm that is proactive, scalable and robust. In this paper, we present a framework for end-to-end proactive management of global heterogeneous networks and their applications. Our framework consists of a three level hierarchy: Network and Protocol Management (NPM) Management Computing System (MCS), and Application Centric Management (ACM). The NPM layer addresses the issues of utilizing existing network management tools, abstracting their collected management information, and providing proactive management services to manage networks and protocols. The MCS layer addresses the core system management issues and provides system management services to enable the development of efficient proactive management of a wide range of network applications. The ACM layer addresses the issues required to develop application specific management techniques, and manage applications so they can meet their requirements in real-time.

The requirements of grand challenge problems and the deployment of gigabit networks makes the network computing framework an attractive and cost effective computing environment with which to interconnect geographically distributed processing and storage resources. Our project, Virtual Distributed Computing Environment (VDCE), provides a problem-solving environment for high-performance distributed computing over wide area networks. VDCE delivers well-defined library functions that relieve end-users of tedious task implementations and also support reusability. In this paper we present the conceptual design of VDCE software architecture, which is defined in three modules: (a) the Application Editor, a user-friendly application development environment that generates the Application Flow Graph (AFG) of an application; (b) the Application Scheduler, which provides an efficient task-to-resource mapping of AFG; and (c) the VDCE Runtime System, which is responsible for running and managing application execution and monitoring the VDCE resources.

Current advances in processor technology, and the rapid development of high speed networking technology, such as ATM, have made high performance network computing an attractive computing environment for large-scale high performance distributed computing (HPDC) applications. However, due to the communications overhead at the host-network interface, most of the HPDC applications are not getting the full benefit of high speed communication networks. This overhead can be attributed to the high cost of operating system calls, context switching, the use of inefficient communication protocols, and the coupling of data and control paths. We present an architecture and implementation for a low-latency, high-throughput message passing tool, that we refer to as the NYNET (ATM wide area network testbed in New York state) Communication System (NCS), which can support a variety of HPDC applications with different Quality of Services (QOS) requirements. NCS uses multithreading to provide efficient techniques that overlap computation and communication. NCS uses read/write trap routines to bypass traditional operating system calls. This reduces latency and avoids using inefficient communication protocols. By separating data and control paths, NCS eliminates unnecessary control transfers. This optimizes the data path and improves performance. Benchmarking results show that the performance of NCS is at least a factor of two better than the performance of corresponding p4 and PVM primitives.