Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Workload”

Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 757 records · Page 42

Multi-core processing and scheduling performance in CMS

Commodity hardware is going many-core. We might soon not be able to satisfy the job memory needs per core in the current single-core processing model in High Energy Physics. In addition, an ever increasing number of independent and incoherent jobs running on the same physical hardware not sharing resources might significantly affect processing performance. It will be essential to effectively utilize the multi-core architecture. CMS has incorporated support for multi-core processing in the event processing framework and the workload management system. Multi-core processing jobs share common data in memory, such us the code libraries, detector geometry and conditions data, resulting in a much lower memory usage than standard single-core independent jobs. Exploiting this new processing model requires a new model in computing resource allocation, departing from the standard single-core allocation for a job. The experiment job management system needs to have control over a larger quantum of resource since multi-core aware jobs require the scheduling of multiples cores simultaneously. CMS is exploring the approach of using whole nodes as unit in the workload management system where all cores of a node are allocated to a multi-core job. Whole-node scheduling allows for optimization of the data/workflow management (e.g. I/O caching, local merging) but efficient utilization of all scheduled cores is challenging. Dedicated whole-node queues have been setup at all Tier-1 centers for exploring multi-core processing workflows in CMS. We present the evaluation of the performance scheduling and executing multi-core workflows in whole-node queues compared to the standard single-core processing workflows.

Hernandez, J. M.↗

Integration of Titan supercomputer at OLCF with ATLAS Production System

The PanDA (Production and Distributed Analysis) workload management system was developed to meet the scale and complexity of distributed computing for the ATLAS experiment. PanDA managed resources are distributed worldwide, on hundreds of computing sites, with thousands of physicists accessing hundreds of Petabytes of data and the rate of data processing already exceeds Exabyte per year. While PanDA currently uses more than 200,000 cores at well over 100 Grid sites, future LHC data taking runs will require more resources than Grid computing can possibly provide. Additional computing and storage resources are required. Therefore ATLAS is engaged in an ambitious program to expand the current computing model to include additional resources such as the opportunistic use of supercomputers. In this paper we will describe a project aimed at integration of ATLAS Production System with Titan supercomputer at Oak Ridge Leadership Computing Facility (OLCF). Current approach utilizes modified PanDA Pilot framework for job submission to Titan's batch queues and local data management, with lightweight MPI wrappers to run single node workloads in parallel on Titan's multi-core worker nodes. It provides for running of standard ATLAS production jobs on unused resources (backfill) on Titan. The system already allowed ATLAS to collect on Titan millions of core-hours per month, execute hundreds of thousands jobs, while simultaneously improving Titans utilization efficiency. We will discuss the details of the implementation, current experience with running the system, as well as future plans aimed at improvements in scalability and efficiency.

97 MATHEMATICS AND COMPUTING↗

Empirically-calibrated H100 node power models for accurate AI training energy estimation

Accurately quantifying the energy use of artificial intelligence (AI) training is critical for infrastructure planning, carbon accounting, and sustainable data center operation, but few studies have directly measured the power consumption of production workloads on contemporary hardware. By combining empirical measurements from Brookhaven National Laboratory during AI training on 8-graphics-processing-unit H100 systems with open-source benchmarking data, we develop statistical models relating computational intensity to node-level power consumption. We measure the gap between manufacturer-rated thermal design power (TDP) and actual power demand during AI training. Our analysis reveals that even computationally intensive workloads operate at only 76% of the 10.2 kW TDP rating. Our architecture-specific model, calibrated to floating-point operations, predicts energy consumption with 11.4% mean absolute percentage error, significantly outperforming TDP-based approaches (27%–37% error). We identified distinct power signatures between transformer and convolutional neural network architectures, with transformers showing characteristic fluctuations that may impact grid stability. These results provide a measurement-grounded basis for improving AI training energy estimates, enabling more reliable infrastructure sizing, cost projections, and environmental impact assessments.

Newkirk, Alex C↗

Exploring Hardware Support For Scaling Irregular Applications on Multi-node Multi-core Architectures

With the recent emergence of large-scale knowledge dis- covery, data mining and social network analysis, irregular applications have gained renewed interest. Classic cache-based high-performance architectures do not provide optimal performances with such kind of workloads, mainly due to the very low spatial and temporal locality of the irregular control and memory access patterns. In this paper, we present a multi-node, multi-core, fine-grained multi-threaded shared-memory system architecture specifically designed for the execution of large-scale irregular applications, and built on top of three pillars, that we believe are fundamental to support these workloads. First, we offer transparent hardware support for Partitioned Global Address Space (PGAS) to provide a large globally-shared address space with no software library overhead. Second, we employ multi-threaded multi-core processing nodes to achieve the necessary latency tolerance required by accessing global memory, which potentially resides in a remote node. Finally, we devise hardware support for inter-thread synchronization on the whole global address space. We first model the performances by using an analytical model that takes into account the main architecture and application characteristics. We describe the hardware design of the proposed cus- tom architectural building blocks that provide support for the above- mentioned three pillars. Finally, we present a limited-scale evaluation of the system on a multi-board FPGA prototype with typical irregular kernels and benchmarks. The experimental evaluation demonstrates the architecture performance scalability for different configurations of the whole system.

FPGA↗

A Runtime Environment for Supporting Research in Resilient HPC System Software & Tools

The high-performance computing (HPC) community continues to increase the size and complexity of hardware platforms that support advanced scientific workloads. The runtime environment (RTE) is a crucial layer in the software stack for these large-scale systems. The RTE manages the interface between the operating system and the application running in parallel on the machine. The deployment of applications and tools on large-scale HPC computing systems requires the RTE to manage process creation in a scalable manner, support sparse connectivity, and provide fault tolerance. We have developed a new RTE that provides a basis for building distributed execution environments and developing tools for HPC to aid research in system software and resilience. This paper describes the software architecture of the Scalable runTime Component Infrastructure (STCI), which is intended to provide a complete infrastructure for scalable start-up and management of many processes in large-scale HPC systems. We highlight features of the current implementation, which is provided as a system library that allows developers to easily use and integrate STCI in their tools and/or applications. The motivation for this work has been to support ongoing research activities in fault-tolerance for large-scale systems. We discuss the advantages of the modular framework employed and describe two use cases that demonstrate its capabilities: (i) an alternate runtime for a Message Passing Interface (MPI) stack, and (ii) a distributed control and communication substrate for a fault-injection tool.

Computer Science↗

Containers for Massive Ensemble of I/O Bound Hierarchical Coupled Simulations

We present our experience using containers to scale up a massive ensemble of coupled I/O bound workloads on the NERSC Cori supercomputer. We describe the design of a hierarchical simulation structure using the Integrated Plasma Simulator (IPS) that enables the flexible execution of coupled simulations at the system, node, and core level using the same coupling abstraction and API. The hierarchical design allows for the node-level execution to be efficiently executed using containers while not impacting the structure of the simulation at the system level. We demonstrate the viability of the approach by presenting experimental results from applications in coupled fusion plasma simulations that illustrate the performance impact of using containers to deploy the node-level workloads, in conjunction with the user mountable XFS file systems to ameliorate the load on the Lustre parallel file system. We also present results from production runs showing the ability of the ensemble simulations to scale to hundreds of Cori Haswell nodes, with little or no overhead.

Elwasif, Wael↗

Overlapped checkpointing with hardware assist

We present a new approach to handling the demanding I/O workload incurred during checkpoint writes encountered in High Performance Computing. Prior efforts to improve performance have been primarily bound by mechanical limitations of the hard drive. Our research surpasses this limitation by providing a method to: (1) write checkpoint data to a high-speed, non-volatile buffer, and (2) asynchronously write this data to permanent storage while resuming computation. This removes the hard drive from the critical data path because our I/O node based buffers isolate the compute nodes from the storage servers. This solution is feasible because of industry declines in cost for high-capacity, non-volatile storage technologies. Testing was conducted on a small-scale cluster to prove the design, and then scaled at Los Alamos National Laboratory. Results show a definitive speedup factor for select workloads over writing directly to a typical global parallel file system; the Panasas ActiveScale File System.

97 MATHEMATICS AND COMPUTING↗

On the Feasibility of Dynamic Power Steering

While high performance has always been the primary constraint behind large-scale system design, future systems will be built with increasing energy efficiency in mind. Mechanisms such as fine-grained power scaling and gating will provide tools to system-software and application developers to ensure the most efficient use of tightly constrained power budgets. Such approaches to-date have been focused on node level optimizations to impact overall system energy efficiency. In this work we introduce Dynamic Power Steering, in which power can be dynamically routed across a system to resources where it will be of most benefit and away from other resources to maintain a near-constant overall power budget. This, a higher level algorithmic approach to improving energy efficiency, considers the whole extent of a system being used by an application. It can be used for applications in which there is load imbalance that varies over its execution. Using two classes of applications, namely those that contain a wave front type processing, and a particle-in-cell, we quantify the benefit of Dynamic Power Steering for a variety of workload characteristics and derive some insight into the ways in which workload behavior affect Power Steering applicability.

Barker, Kevin J.↗

GCoD: Graph Convolutional Network Acceleration via Dedicated Algorithm and Accelerator Co-Design

Graph Convolutional Networks (GCNs) have emerged as the state-of-the-art graph learning model. However, it remains notoriously challenging to inference GCNs over large graph datasets, limiting their application to large real-world graphs and hindering the exploration of deeper and more sophisticated GCN graphs. This is because real-world graphs can be extremely large and sparse. Furthermore, the node degree of GCNs tends to follow the power-law distribution and therefore have highly irregular adjacency matrices, resulting in prohibitive inefficiencies in both data processing and movement and thus substantially limiting the achievable GCN acceleration efficiency. To this end, this paper proposes the first GCN algorithm and accelerator Co-Design framework dubbed GCoD which can largely alleviate the aforementioned GCN irregularity and boost GCNs' inference efficiency. Specifically, on the algorithm level, GCoD integrates a divide and conquer GCN training strategy that polarizes the graphs to be either denser or sparser in local neighborhoods without compromising the model accuracy, resulting in graph adjacency matrices that (mostly) have merely two levels of workload and enjoys largely enhanced regularity and thus ease of acceleration. On the hardware level, we further develop a dedicated two-pronged accelerator with a separated engine to process each of the aforementioned workloads, further boosting the overall utilization and acceleration efficiency. Extensive experiments and ablation studies validate that our GCoD consistently outperforms state-of-the-art designs in terms of accelerator efficiency while maintaining or even improving the task accuracy. Additionally, we visualize GCoD trained graph adjacency matrices to better understand its advantages. All codes and pre-trained models will be released upon acceptance.

You, Haoran↗

JACC.shared: Leveraging HPC Metaprogramming and Performance Portability for Computations That Use Shared Memory GPUs

In this work, we present JACC.shared, a new feature of Julia for ACCelerators (JACC), which is the performanceportable and metaprogramming model of the just-in-time and LLVM-based Julia language. This new feature allows JACC applications to leverage the high-performance computing (HPC) capabilities of high-bandwidth, on-chip GPU memory. Historically, exploiting high-bandwidth, shared-memory GPUs has not been a priority for high-level programming solutions. JACC.shared covers that gap for the first time, thereby providing a highlevel, portable, and easy-to-use solution for programmers to exploit this memory and supporting all current major accelerator architectures. Well-known HPC and AI workloads, such as multi/hyperspectral imaging and AI convolutions, have been used to evaluate JACC.shared on two exascale GPU architectures hosted by some of the most powerful US Department of Energy supercomputers: Perlmutter (NVIDIA A100) and Frontier (AMD MI250X). The performance evaluation reports speedup of up to 3.5× by adding only one line of code to the base codes, thus providing important accelerators in a simple, portable, and transparent way and elevating the programming productivity and performance-portability capabilities for Julia/JACC HPC, AI, and scientific applications.

Valero Lara, Pedro [ORNL] (ORCID:0000000214794310)↗

Dynamic Load Balancing on Single- and Multi-GPU Systems

The computational power provided by many-core graphics processing units (GPUs) has been exploited in many applications. The programming techniques supported and employed on these GPUs are not sufficient to address problems exhibiting irregular, and unbalanced workload. The problem is exacerbated when trying to effectively exploit multiple GPUs, which are commonly available in many modern systems. In this paper, we propose a task-based dynamic load-balancing solution for single- and multi-GPU systems. The solution allows load balancing at a finer granularity than what is supported in existing APIs such as NVIDIA’s CUDA. We evaluate our approach using both micro-benchmarks and a molecular dynamics application that exhibits significant load imbalance. Experimental results with a single-GPU configuration show that our fine-grained task solution can utilize the hardware more efficiently than the CUDA scheduler for unbalanced workload. On multi-GPU systems, our solution achieves near-linear speedup, load balance, and significant performance improvement over techniques based on standard CUDA APIs.

97 MATHEMATICS AND COMPUTING↗

Time Utility Functions for Modeling and Evaluating Resource Allocations in a Heterogeneous Computing System

This study considers a heterogeneous computing system and corresponding workload being investigated by the Extreme Scale Systems Center (ESSC) at Oak Ridge National Laboratory (ORNL). The ESSC is part of a collaborative effort between the Department of Energy (DOE) and the Department of Defense (DoD) to deliver research, tools, software, and technologies that can be integrated, deployed, and used in both DOE and DoD environments. The heterogeneous system and workload described here are representative of a prototypical computing environment being studied as part of this collaboration. Each task can exhibit a time-varying importance or utility to the overall enterprise. In this system, an arriving task has an associated priority and precedence. The priority is used to describe the importance of a task, and precedence is used to describe how soon the task must be executed. These two metrics are combined to create a utility function curve that indicates how valuable it is for the system to complete a task at any given moment. This research focuses on using time-utility functions to generate a metric that can be used to compare the performance of different resource schedulers in a heterogeneous computing system. The contributions of this paper are: (a) a mathematical model of a heterogeneous computing system where tasks arrive dynamically and need to be assigned based on their priority, precedence, utility characteristic class, and task execution type, (b) the use of priority and precedence to generate time-utility functions that describe the value a task has at any given time, (c) the derivation of a metric based on the total utility gained from completing tasks to measure the performance of the computing environment, and (d) a comparison of the performance of resource allocation heuristics in this environment.

99 GENERAL AND MISCELLANEOUS↗

Efficient, Parallel At-scale Correlation Analysis for Atom Probe Tomography on Hybrid Architectures

Atom probe tomography (APT) is a material probing technique that has undergone dramatic improvements in its capability to map individual atoms within a material sample resulting in data files with hundreds of millions of atoms. Understanding the nano-structural features hidden in these massive amounts of atomic data is a crucial analysis task for materials scientists. However, fast analysis capabilities for large APT workloads remains a critical bottleneck. In this paper, we present the design, implementation and detailed performance evaluations of a parallel software capable of efficiently performing extremely time-consuming correlation analyses of massive high density APT data. Starting with shared memory implementations to motivate our design choices, we extend the implementation to hybrid architectures keeping realistic APT workloads in mind. Detailed performance analyses of three different parallel implementations of the software are supported by empirical results on a Cray XC30 and a Cray XC40 architecture. Its usefulness is demonstrated by reducing the turnaround time of an end-to-end APT correlation analysis on 100 millions atoms by three orders of magnitude using 2048 MPI ranks on 1024 nodes (24 cores per node) of a Cray XC30. The software reported here equips material scientists for the first time with a high-speed scalable capability for efficient and timely analyses of massive APT data.

Lu, Hao↗

SYMBIOSYS: A Methodology for Performance Analysis of Composable HPC Data Services

Microservices are a powerful new way of building, customizing, and deploying distributed services owing to their flexibility and maintainability. Several large-scale distributed platforms have emerged to serve the growing needs of data-centric workloads and services in commercial computing. Concurrently, high-performance computing (HPC) systems and software are rapidly evolving to meet the demands of diversified applications and heterogeneity. The interplay of hardware factors, software configuration parameters, and the flexibility offered with a microservice architecture makes it nontrivial to estimate the optimal service instantiation for a given application workload. Further, this problem is exacerbated when considering that these services operate in a dynamic and heterogeneous HPC environment. An optimally integrated service can be vastly more performant than a haphazardly integrated one. Existing performance tools for HPC either fail to understand the request-response model of communication inherent to microservices or they operate within a narrow scope, limiting the insight that can be gleaned from employing them in isolation. We propose a methodology for integrated performance analysis of HPC microservices frameworks and applications called SYMBIOSYS. We describe its design and implementation within the context of the Mochi framework. This integration is achieved by combining distributed callpath profiling and tracing with a performance data exchange strategy that collects fine-grained, low-level metrics from the RPC communication library and network layers. The result is a portable, low-overhead performance analysis setup that provides a holistic profile of the dependencies among microservices and how they interact with the Mochi RPC software stack. Using HEPnOS, a production-quality Mochi data service, we demonstrate the low-overhead operation of SYMBIOSYS at scale and use it to identify the root causes of poorly performing service configurations.

microservices↗

UnifyFS: A User-level Shared File System for Unified Access to Distributed Local Storage

We introduce UnifyFS, a user-level file system that aggregates node-local storage tiers available on high performance computing (HPC) systems and makes them available to HPC applications under a unified namespace. UnifyFS employs transparent I/O interception, so it does not require changes to application code and is compatible with commonly used HPC I/O libraries. The design of UnifyFS supports the predominant HPC I/O workloads and is optimized for bulk-synchronous I/O patterns. Furthermore, UnifyFS provides customizable file system semantics to flexibly adapt its behavior for diverse I/O workloads and storage devices. In this paper, we discuss the unique design goals and architecture of UnifyFS and evaluate its performance on a leadership-class HPC system. In our experimental results, we demonstrate that UnifyFS exhibits excellent scaling performance for write operations and can improve the performance of application checkpoint operations by as much as 3× versus a tuned configuration.

Brim, Michael↗

Preemptive resource management for dynamically arriving tasks in an oversubscribed heterogeneous computing system

We design resource management heuristics that assign serial tasks to the nodes of a heterogeneous high performance computing (HPC) system. The value of completing these tasks is modeled using monotonically decreasing utility functions that represent the time-varying importance of the task. The value of completing a task is equal to its utility function at the time of its completion. The overall performance of this system is measured using the total utility earned by all tasks during some interval of time. To maximize the performance of such a system where the preemption of tasks is possible, we have designed, analyzed, and compared a set of resource allocation heuristic techniques. We combine two utility-aware heuristics with three different preemption techniques to create six preemption-capable heuristics. We also consider the two utility-aware heuristics without preemption. We use simulation studies to evaluate this set of eight heuristics and compare them with an FCFS heuristic, which is often used in real systems, and random assignments. In general, our set of eight heuristics is able to significantly outperform the comparison heuristics, and the preemption-capable heuristics are able to significantly increase the utility earned compared to the heuristics that do not use preemption. We analyze the performance tradeoffs among the different preemption-capable heuristics under a variety of oversubscribed workload environments.

97 MATHEMATICS AND COMPUTING↗

XaaS: Acceleration as a Service to Enable Productive High-Performance Cloud Computing

High-performance computing (HPC) and the cloud have evolved independently, specializing their innovations into performance or productivity. Acceleration as a Service (XaaS) is a recipe to empower both fields with a shared execution platform that provides transparent access to computing resources, regardless of the underlying cloud or HPC service provider. Bridging HPC and cloud advancements, XaaS presents a unified architecture built on performance-portable containers. Here, our converged model concentrates on low-overhead, high-performance communication and computing, targeting resource-intensive workloads from climate simulations to machine learning. XaaS lifts the restricted allocation model of Function as a Service (FaaS), allowing users to benefit from the flexibility and efficient resource utilization of serverless computing while supporting long-running and performance-sensitive workloads from HPC.

97 MATHEMATICS AND COMPUTING↗

Converged Computing: A Best of Both Worlds of High-Performance Computing and Cloud

Collaboration between Cloud and High Performance Computing (HPC) communities has accelerated in the last half decade. A common goal to run batch workloads combined with a desire for reproducibility, automation, and optimization has led to successful projects that range from container technologies to workload management and security. This span of current and future work defines a novel “Converged Computing” paradigm that aims to combine the best of both worlds, both from a technological and cultural standpoint. Furthermore, in this Special Issue, we review common themes in the space, showcasing current work and encouraging a continued effort toward innovative ideas that will enable the next generations of scientific discovery.

97 MATHEMATICS AND COMPUTING↗