arXiv subjects

Michael E. Papka

Publications and source records attributed to Michael E. Papka.

47 records · Page 3Linked to original sources

arXiv2023

DeepSpeed4Science Initiative: Enabling Large-Scale Scientific Discovery through Sophisticated AI System Technologies

In the upcoming decade, deep learning may revolutionize the natural sciences, enhancing our capacity to model and predict natural occurrences. This could herald a new era of scientific exploration, bringing significant advancements across sectors from drug development to renewable energy. To answer this call, we present DeepSpeed4Science initiative (deepspeed4science.ai) which aims to build unique capabilities through AI system technology innovations to help domain experts to unlock today's biggest science mysteries. By leveraging DeepSpeed's current technology pillars (training, inference and compression) as base technology enablers, DeepSpeed4Science will create a new set of AI system technologies tailored for accelerating scientific discoveries by addressing their unique complexity beyond the common technical approaches used for accelerating generic large language models (LLMs). In this paper, we showcase the early progress we made with DeepSpeed4Science in addressing two of the critical system challenges in structural biology research.

cs.AI↗

arXiv2023

A Comprehensive Performance Study of Large Language Models on Novel AI Accelerators

Artificial intelligence (AI) methods have become critical in scientific applications to help accelerate scientific discovery. Large language models (LLMs) are being considered as a promising approach to address some of the challenging problems because of their superior generalization capabilities across domains. The effectiveness of the models and the accuracy of the applications is contingent upon their efficient execution on the underlying hardware infrastructure. Specialized AI accelerator hardware systems have recently become available for accelerating AI applications. However, the comparative performance of these AI accelerators on large language models has not been previously studied. In this paper, we systematically study LLMs on multiple AI accelerators and GPUs and evaluate their performance characteristics for these models. We evaluate these systems with (i) a micro-benchmark using a core transformer block, (ii) a GPT- 2 model, and (iii) an LLM-driven science use case, GenSLM. We present our findings and analyses of the models' performance to better understand the intrinsic capabilities of AI accelerators. Furthermore, our analysis takes into account key factors such as sequence lengths, scaling behavior, sparsity, and sensitivity to gradient accumulation steps.

cs.PF↗

arXiv2023

A Multi-Level, Multi-Scale Visual Analytics Approach to Assessment of Multifidelity HPC Systems

The ability to monitor and interpret of hardware system events and behaviors are crucial to improving the robustness and reliability of these systems, especially in a supercomputing facility. The growing complexity and scale of these systems demand an increase in monitoring data collected at multiple fidelity levels and varying temporal resolutions. In this work, we aim to build a holistic analytical system that helps make sense of such massive data, mainly the hardware logs, job logs, and environment logs collected from disparate subsystems and components of a supercomputer system. This end-to-end log analysis system, coupled with visual analytics support, allows users to glean and promptly extract supercomputer usage and error patterns at varying temporal and spatial resolutions. We use multiresolution dynamic mode decomposition (mrDMD), a technique that depicts high-dimensional data as correlated spatial-temporal variations patterns or modes, to extract variation patterns isolated at specified frequencies. Our improvements to the mrDMD algorithm help promptly reveal useful information in the massive environment log dataset, which is then associated with the processed hardware and job log datasets using our visual analytics system. Furthermore, our system can identify the usage and error patterns filtered at user, project, and subcomponent levels. We exemplify the effectiveness of our approach with two use scenarios with the Cray XC40 supercomputer.

cs.HC↗

arXiv2022

Linking Scientific Instruments and HPC: Patterns, Technologies, Experiences

Powerful detectors at modern experimental facilities routinely collect data at multiple GB/s. Online analysis methods are needed to enable the collection of only interesting subsets of such massive data streams, such as by explicitly discarding some data elements or by directing instruments to relevant areas of experimental space. Such online analyses require methods for configuring and running high-performance distributed computing pipelines--what we call flows--linking instruments, HPC (e.g., for analysis, simulation, AI model training), edge computing (for analysis), data stores, metadata catalogs, and high-speed networks. In this article, we review common patterns associated with such flows and describe methods for instantiating those patterns. We also present experiences with the application of these methods to the processing of data from five different scientific instruments, each of which engages HPC resources for data inversion, machine learning model training, or other purposes. We also discuss implications of these new methods for operators and users of scientific facilities.

cs.DC↗

arXiv2021

Toward Real-time Analysis of Experimental Science Workloads on Geographically Distributed Supercomputers

Massive upgrades to science infrastructure are driving data velocities upwards while stimulating adoption of increasingly data-intensive analytics. While next-generation exascale supercomputers promise strong support for I/O-intensive workflows, HPC remains largely untapped by live experiments, because data transfers and disparate batch-queueing policies are prohibitive when faced with scarce instrument time. To bridge this divide, we introduce Balsam: a distributed orchestration platform enabling workflows at the edge to securely and efficiently trigger analytics tasks across a user-managed federation of HPC execution sites. We describe the architecture of the Balsam service, which provides a workflow management API, and distributed sites that provision resources and schedule scalable, fault-tolerant execution. We demonstrate Balsam in efficiently scaling real-time analytics from two DOE light sources simultaneously onto three supercomputers (Theta, Summit, and Cori), while maintaining low overheads for on-demand computing, and providing a Python library for seamless integration with existing ecosystems of data analysis tools.

cs.DC↗

arXiv2021

BFTrainer: Low-Cost Training of Neural Networks on Unfillable Supercomputer Nodes

Supercomputer FCFS-based scheduling policies result in many transient idle nodes, a phenomenon that is only partially alleviated by backfill scheduling methods that promote small jobs to run before large jobs. Here we describe how to realize a novel use for these otherwise wasted resources, namely, deep neural network (DNN) training. This important workload is easily organized as many small fragments that can be configured dynamically to fit essentially any node*time hole in a supercomputer's schedule. We describe how the task of rescaling suitable DNN training tasks to fit dynamically changing holes can be formulated as a deterministic mixed integer linear programming (MILP)-based resource allocation algorithm, and show that this MILP problem can be solved efficiently at run time. We show further how this MILP problem can be adapted to optimize for administrator- or user-defined metrics. We validate our method with supercomputer scheduler logs and different DNN training scenarios, and demonstrate efficiencies of up to 93% compared with running the same training tasks on dedicated nodes. Our method thus enables substantial supercomputer resources to be allocated to DNN training with no impact on other applications.

cs.DC↗

arXiv2021

Deep Reinforcement Agent for Scheduling in HPC

Cluster scheduler is crucial in high-performance computing (HPC). It determines when and which user jobs should be allocated to available system resources. Existing cluster scheduling heuristics are developed by human experts based on their experience with specific HPC systems and workloads. However, the increasing complexity of computing systems and the highly dynamic nature of application workloads have placed tremendous burden on manually designed and tuned scheduling heuristics. More aggressive optimization and automation are needed for cluster scheduling in HPC. In this work, we present an automated HPC scheduling agent named DRAS (Deep Reinforcement Agent for Scheduling) by leveraging deep reinforcement learning. DRAS is built on a novel, hierarchical neural network incorporating special HPC scheduling features such as resource reservation and backfilling. A unique training strategy is presented to enable DRAS to rapidly learn the target environment. Once being provided a specific scheduling objective given by system manager, DRAS automatically learns to improve its policy through interaction with the scheduling environment and dynamically adjusts its policy as workload changes. The experiments with different production workloads demonstrate that DRAS outperforms the existing heuristic and optimization approaches by up to 45%.

cs.DC↗

arXiv2020

Scheduling Beyond CPUs for HPC

High performance computing (HPC) is undergoing significant changes. The emerging HPC applications comprise both compute- and data-intensive applications. To meet the intense I/O demand from emerging data-intensive applications, burst buffers are deployed in production systems. Existing HPC schedulers are mainly CPU-centric. The extreme heterogeneity of hardware devices, combined with workload changes, forces the schedulers to consider multiple resources (e.g., burst buffers) beyond CPUs, in decision making. In this study, we present a multi-resource scheduling scheme named BBSched that schedules user jobs based on not only their CPU requirements, but also other schedulable resources such as burst buffer. BBSched formulates the scheduling problem into a multi-objective optimization (MOO) problem and rapidly solves the problem using a multi-objective genetic algorithm. The multiple solutions generated by BBSched enables system managers to explore potential tradeoffs among various resources, and therefore obtains better utilization of all the resources. The trace-driven simulations with real system workloads demonstrate that BBSched improves scheduling performance by up to 41% compared to existing methods, indicating that explicitly optimizing multiple resources beyond CPUs is essential for HPC scheduling.

cs.DC↗

arXiv2019

Balsam: Automated Scheduling and Execution of Dynamic, Data-Intensive HPC Workflows

We introduce the Balsam service to manage high-throughput task scheduling and execution on supercomputing systems. Balsam allows users to populate a task database with a variety of tasks ranging from simple independent tasks to dynamic multi-task workflows. With abstractions for the local resource scheduler and MPI environment, Balsam dynamically packages tasks into ensemble jobs and manages their scheduling lifecycle. The ensembles execute in a pilot "launcher" which (i) ensures concurrent, load-balanced execution of arbitrary serial and parallel programs with heterogeneous processor requirements, (ii) requires no modification of user applications, (iii) is tolerant of task-level faults and provides several options for error recovery, (iv) stores provenance data (e.g task history, error logs) in the database, (v) supports dynamic workflows, in which tasks are created or killed at runtime. Here, we present the design and Python implementation of the Balsam service and launcher. The efficacy of this system is illustrated using two case studies: hyperparameter optimization of deep neural networks, and high-throughput single-point quantum chemistry calculations. We find that the unique combination of flexible job-packing and automated scheduling with dynamic (pilot-managed) execution facilitates excellent resource utilization. The scripting overheads typically needed to manage resources and launch workflows on supercomputers are substantially reduced, accelerating workflow development and execution.

cs.DC↗

arXiv2017

Predicting Research that will be Cited in Policy Documents

Scientific publications and other genres of research output are increasingly being cited in policy documents. Citations in documents of this nature could be considered a critical indicator of the significance and societal impact of the research output. In this study, we built classification models that predict whether a particular research work is likely to be cited in a public policy document based on the attention it received online, primarily on social media platforms. We evaluated the classifiers based on their accuracy, precision, and recall values. We found that Random Forest and Multinomial Naive Bayes classifiers performed better overall.

cs.DL↗

arXiv2011

Multiscale simulation of blood flow in brain arteries with an aneurysm

Interfacing atomistic-based with continuum-based simulation codes is now required in many multiscale physical and biological systems. We present the first results from coupled atomistic-continuum simulations on 190,000 processors. Platelet aggregation in the patient-specific model of an aneurysm has been modeled using a high-order spectral/hp element Navier-Stokes solver with a stochastic (coarse-grained) Molecular Dynamics solver based on Dissipative Particle Dynamics (DPD).

physics.flu-dyn↗