Engineering Papers⌕ Search

Engineering topics

Taylor, Valerie

Publications and source records attributed to Taylor, Valerie.

14 records
All sourcesNASA NTRSDOE OSTI
DOE OSTI2025

LLM-Inference-Bench: Inference Benchmarking of Large Language Models on AI Accelerators

Large Language Models (LLMs) have propelled groundbreaking advancements across several domains and are commonly used for text generation applications. However, the computational demands of these complex models pose significant challenges, requiring efficient hardware acceleration. Benchmarking the performance of LLMs across diverse hardware platforms is crucial to understanding their scalability and throughput characteristics. We introduce LLM-Inference-Bench, a comprehensive benchmarking suite to evaluate the hardware inference performance of LLMs. We thoroughly analyze diverse hardware platforms, including GPUs from Nvidia and AMD and specialized AI accelerators, Intel Habana and SambaNova. Our evaluation includes several LLM inference frameworks and models from LLaMA, Mistral, and Qwen families with 7B and 70B parameters. Our benchmarking results reveal the strengths and limitations of various models, hardware platforms, and inference frameworks. We provide an interactive dashboard to help identify configurations for optimal performance for a given hardware platform.

Chitty-Venkata, Krishna Teja↗
DOE OSTI2024

ytopt: Autotuning Scientific Applications for Energy Efficiency at Large Scales

As we enter the exascale computing era, efficiently utilizing power and optimizing the performance of scientific applications under power and energy constraints has become critical and challenging. We propose a low-overhead autotuning framework to autotune performance and energy for various hybrid MPI/OpenMP scientific applications at large scales and to explore the tradeoffs between application runtime and power/energy for energy efficient application execution, then use this framework to autotune four ECP proxy applications—XSBench, AMG, SWFFT, and SW4lite. Our approach uses Bayesian optimization with a Random Forest surrogate model to effectively search parameter spaces with up to 6 million different configurations on two large-scale HPC production systems, Theta at Argonne National Laboratory and Summit at Oak Ridge National Laboratory. The experimental results show that our autotuning framework at large scales has low overhead and achieves good scalability. Using the proposed autotuning framework to identify the best configurations, we achieve up to 91.59% performance improvement, up to 21.2% energy savings, and up to 37.84% EDP (energy delay product) improvement on up to 4096 nodes.

Autotuning↗
DOE OSTI2024

Integrating ytopt and libEnsemble to autotune OpenMC

Ytopt is a Python machine-learning-based autotuning software package developed within the ECP PROTEAS-TUNE project. The ytopt software adopts an asynchronous search framework that consists of sampling a small number of input parameter configurations and progressively fitting a surrogate model over the input-output space until exhausting the user-defined maximum number of evaluations or the wall-clock time. libEnsemble is a Python toolkit for coordinating workflows of asynchronous and dynamic ensembles of calculations across massively parallel resources developed within the ECP PETSc/TAO project. libEnsemble helps users take advantage of massively parallel resources to solve design, decision, and inference problems and expands the class of problems that can benefit from increased parallelism. In this paper we present our methodology and framework to integrate ytopt and libEnsemble to take advantage of massively parallel resources to accelerate the autotuning process. Specifically, we focus on using the proposed framework to autotune the ECP ExaSMR application OpenMC, an open source Monte Carlo particle transport code. OpenMC has seven tunable parameters some of which have large ranges such as the number of particles in-flight, which is in the range of 100,000 to 8 million, with its default setting of 1 million. Setting the proper combination of these parameter values to achieve the best performance is extremely time-consuming. Therefore, we apply the proposed framework to autotune the MPI/OpenMP offload version of OpenMC based on a user-defined metric such as the figure of merit (FoM) (particles/s) or energy efficiency energy-delay product (EDP) on Crusher at Oak Ridge Leadership Computing Facility. In conclusion, the experimental results show that we achieve the improvement up to 29.49% in FoM and up to 30.44% in EDP.

Autotuning↗
DOE OSTI2023

Advanced Research Directions on AI for Science, Energy, and Security: Report on Summer 2022 Workshops

Over the past decade, fundamental changes in artificial intelligence (AI)—from foundational to applied—have delivered dramatic insights across a wide breadth of U.S. Department of Energy (DOE) mission space. AI is helping to augment and improve scientific and engineering workflows (e.g., for control, design, and dramatic performance gains through surrogate models) in national security, the Office of Science, and DOE’s applied energy programs. The progress and potential for AI in DOE science was captured in the 2020 “AI for Science” report from the DOE laboratory community in collaboration with academia and industry. Specific scientific areas ready to further leverage the power of AI ranged from the scale and performance of computational models to data analysis to creating new classes of observations using computer vision. Since that report, the scale and scope of scientific AI have accelerated, revealing new, emergent properties that yield insights that go beyond enabling opportunities to being potentially transformative in the way that scientific problems are posed and solved. Thus, under the guidance of both the Office of Science (SC) and the National Nuclear Security Administration (NNSA), the DOE national laboratories organized a series of workshops in 2022 to gather input on new and rapidly emerging opportunities and challenges of scientific AI. This 2023 report is a synthesis of those workshops. The scientific community believes AI can have a foundational impact on a broad range of DOE missions, including science, energy, and national security. Further, DOE has unique capabilities that enable the community to drive progress in scientific use of AI, building on long-standing DOE strengths and investments in computation, data, and communications infrastructure, spanning the Energy Sciences Network (ESnet), the Exascale Computing Project (ECP), and integrative programs such as the NNSA Office of Defense Programs Advanced Simulation and Computing (ASC) and the SC Scientific Discovery through Advanced Computing (SciDAC) programs.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

2022 AI Testbed Expeditions Report

By exploiting the coherent properties of a light source, coherent diffraction imaging (CDI) is able to obtain the sample image at a nanoscale resolution using the measured diffraction pattern. Bragg Coherent Diffraction Imaging (BCDI) has become valuable for recovering the displacement and strain field of crystals, providing a valuable tool in material science and solid-state physics. X-ray ptychography is another emerging CDI technique that can produce a high-resolution image of the extended sample and has become popular in many research areas (e.g., materials science, biology, electronics, and optics characterization). CDI including BCDI and ptychography has become an established technique in Synchrotron Facilities including the Advanced Photon Source (APS) and will greatly benefit from the 100x coherent flux increase of the upcoming APS Upgrade (APSU). The current image formation process in CDI employs iterative phase retrieval algorithms, which is a time-consuming and computationally expensive process. Especially after APSU, the traditional iterative methods will not be able to match the experimental data acquisition speed. We employ deep learning (DL) approach to replace the iterative approaches, therefore allowing hundreds of times faster recovery of the object. We developed AutoPhaseNN, a DL-based approach which learns to solve the inverse problem without labeled data. Taking 3D BCDI as a representative technique, AutoPhaseNN has been demonstrated to be one hundred times faster than traditional iterative phase retrieval methods while providing comparable image quality. The current network is trained with 64 x 64 x 64 data size, to achieve higher resolution imaging, we will need to scale the network to input and train/infer 3D arrays of size 256 x 256 x 256 (today) and of size 2560x2560x2560 (APSU). However, the scalability of the network is restricted due to the memory-intensive training process. To perform the training for a 256 x 256 x 256 data size, the required memory exceeds the capacity of the current machine. In this project, we explore using Sambanova system to train the network for the direct data inversion for CDI.

36 MATERIALS SCIENCE↗
DOE OSTI2022

AI@DOE Interim Executive Report

Interim executive report for DOE’s Office of Science, NNSA, and Applied Energy Offices, in collaboration with the Artificial Intelligence and Technology Office, DOE AI roundtable workshops (December 2021 through February 2022) (“AI@DOE”). These workshops were used to help identify AI research requirements and priorities for the next decade for Department of Energy in developing machine-learning-based prediction and decision-support capabilities that address and anticipate DOE mission challenges. These missions’ challenges for AI include: AI to advance scientific discovery and technological innovation AI to support high consequence decisions while managing risk AI to improve DOE’s responsiveness to national & global issues AI to assist stewardship of the environment and national critical infrastructure These will be addressed in large part through the development and application of new and powerful artificial intelligence and/or machine learning algorithms and strategies. https://web.cvent.com/event/fc3922f8-fc75-4041-a317-f13a1da44f7c/summary?locale=en-US&i=2Wx2tjbStkydRFq_N2whPw

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Performance and power modeling and prediction using MuMMI and 10 machine learning methods

Energy-efficient scientific applications require insight into how high performance computing system features impact the applications' power and performance. This insight can result from the development of performance and power models. Here, in this article, we use the modeling and prediction tool MuMMI (Multiple Metrics Modeling Infrastructure) and 10 machine learning methods to model and predict performance and power consumption and compare their prediction error rates. We use an algorithm-based fault-tolerant linear algebra code and a multilevel checkpointing fault-tolerant heat distribution code to conduct our modeling and prediction study on the Cray XC40 Theta and IBM BG/Q Mira at Argonne National Laboratory and the Intel Haswell cluster Shepard at Sandia National Laboratories. Our experimental results show that the prediction error rates in performance and power using MuMMI are less than 10% for most cases. By utilizing the models for runtime, node power, CPU power, and memory power, we identify the most significant performance counters for potential application optimizations, and we predict theoretical outcomes of the optimizations. Based on two collected datasets, we analyze and compare the prediction accuracy in performance and power consumption using MuMMI and 10 machine learning methods.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Hands-On Computer Science: The Array of Things Experimental Urban Instrument

Chicago's Array of Things (AoT) project is aptly described as a technology experiment or a "smart city" prototype. The concept of such an extensible "instrument" arose within a larger translational research vision applying computer science and engineering research for the multidimensional benefit of people and communities in cities. The AoT project hypothesized that wireless intelligent sensor networks could enable both quantitative social science and urban monitoring while also stimulating youth interest in science and technology. Successful deployment of such sensor networks could provide open data from urban measurements not only in support of diverse research questions-in environmental dynamics, urban architecture, engineering, and social sciences-but also informing community groups and city planners. Further, the AoT project and its successor SAGE project are a computer science and engineering experiment, but its success is inextricably tied to community engagement and experiential education. Simply put, community acceptance is a prerequisite to installing and testing the instrument.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2021

Autotuning PolyBench benchmarks with LLVM Clang/Polly loop optimization pragmas using Bayesian optimization

Here, we develop a ytopt autotuning framework that leverages Bayesian optimization to explore the parameter space search and compare four different supervised learning methods within Bayesian optimization and evaluate their effectiveness. We select six of the most complex PolyBench benchmarks and apply the newly developed LLVM Clang/Polly loop optimization pragmas to the benchmarks to optimize them. We then use the autotuning framework to optimize the pragma parameters to improve their performance. The experimental results show that our autotuning approach outperforms the other compiling methods to provide the smallest execution time for the benchmarks syr2k, 3mm, heat-3d, lu, and covariance with two large datasets in 200 code evaluations for effectively searching the parameter spaces with up to 170,368 different configurations. We find that the Floyd-Warshall benchmark did not benefit from autotuning. To cope with this issue, we provide some compiler option solutions to improve the performance. Then we present loop autotuning without a user's knowledge using a simple mctree autotuning framework to further improve the performance of the Floyd-Warshall benchmark. We also extend the ytopt autotuning framework to tune a deep learning application.

79 ASTRONOMY AND ASTROPHYSICS↗
DOE OSTI2021

Utilizing ensemble learning for performance and power modeling and improvement of parallel cancer deep learning CANDLE benchmarks

Abstract Machine learning (ML) continues to grow in importance across nearly all domains in modeling to learn from data. Often a tradeoff exists between a model's ability to minimize bias and variance. In this article, we utilize ensemble learning to combine linear, nonlinear, and tree‐/rule‐based ML methods to cope with the bias‐variance tradeoff and result in more accurate models. We use the datasets collected for two parallel cancer deep learning CANDLE benchmarks, NT3 and P1B2, to build performance and power models based on hardware performance counters using single‐object and multiple‐objects ensemble learning to identify the most important counters for improvement on the Cray XC40 Theta at Argonne National Laboratory. Based on the insights from these models, we improve the performance and energy of P1B2 and NT3 by optimizing the deep learning environments TensorFlow, Keras, Horovod, and Python under the huge page size of 8 MB. Experimental results show that ensemble learning not only produces more accurate models but also provides more robust performance counter ranking. We achieve up to 61.15% performance improvement and up to 62.58% energy saving for P1B2 and up to 55.81% performance improvement and up to 52.60% energy saving for NT3 on up to 24,576 cores.

Wu, Xingfu↗
DOE OSTI2020

AI for Science: Report on the Department of Energy (DOE) Town Halls on Artificial Intelligence (AI) for Science

The report documents the DOE Town Halls held during 2019 at Argonne National Laboratory, Oak Ridge National Laboratory, Lawrence Berkeley National Laboratory, and in Washington, DC. From July to October 2019, the Argonne, Oak Ridge, and Berkeley National Laboratories hosted a series of four town hall meetings attended by more than 1,000 U.S. scientists and engineers. The goal of the town hall series was to examine scientific opportunities in the areas of artificial intelligence (AI), Big Data, and high-performance computing (HPC) in the next decade, and to capture the big ideas, grand challenges, and next steps to realizing these opportunities. In this report and in the Department of Energy (DOE) laboratory community, we use the term “AI for Science” to broadly represent the next generation of methods and scientific opportunities in computing, including the development and application of AI methods (e.g., machine learning, deep learning, statistical methods, data analytics, automated control, and related areas) to build models from data and to use these models alone or in conjunction with simulation and scalable computing to advance scientific research. The AI for Science town hall discussions focused on capturing the transformational uses of AI that employ HPC and/or data analysis, leveraging data sets from HPC simulations or instruments and user facilities, and addressing scientific challenges unique to DOE user facilities and the agency’s wide-ranging fundamental and applied science enterprise.

36 MATERIALS SCIENCE↗
NASA NTRS2001

A Simple Case Study of a Grid Performance System

This document presents a simple case study of a Grid performance system based on the Grid Monitoring Architecture (GMA) being developed by the Grid Forum Performance Working Group. It describes how the various system components would interact for a very basic monitoring scenario, and is intended to introduce people to the terminology and concepts presented in greater detail in other Working Group documents. We believe that by focusing on the simple case first, working group members can familiarize themselves with terminology and concepts, and productively join in the ongoing discussions of the group. In addition, prototype implementations of this basic scenario can be built to explore the feasibility of the proposed architecture and to expose possible shortcomings. Once the simple case is understood and agreed upon, complexities can be added incrementally as warranted by cases not addressed in the most basic implementation described here. Following the basic performance monitoring scenario discussion, unresolved issues are introduced for future discussion.

Aydt, Ruth↗
NASA NTRS2000

Efficient Use of Distributed Systems for Scientific Applications

Distributed computing has been regarded as the future of high performance computing. Nationwide high speed networks such as vBNS are becoming widely available to interconnect high-speed computers, virtual environments, scientific instruments and large data sets. One of the major issues to be addressed with distributed systems is the development of computational tools that facilitate the efficient execution of parallel applications on such systems. These tools must exploit the heterogeneous resources (networks and compute nodes) in distributed systems. This paper presents a tool, called PART, which addresses this issue for mesh partitioning. PART takes advantage of the following heterogeneous system features: (1) processor speed; (2) number of processors; (3) local network performance; and (4) wide area network performance. Further, different finite element applications under consideration may have different computational complexities, different communication patterns, and different element types, which also must be taken into consideration when partitioning. PART uses parallel simulated annealing to partition the domain, taking into consideration network and processor heterogeneity. The results of using PART for an explicit finite element application executing on two IBM SPs (located at Argonne National Laboratory and the San Diego Supercomputer Center) indicate an increase in efficiency by up to 36% as compared to METIS, a widely used mesh partitioning tool. The input to METIS was modified to take into consideration heterogeneous processor performance; METIS does not take into consideration heterogeneous networks. The execution times for these applications were reduced by up to 30% as compared to METIS. These results are given in Figure 1 for four irregular meshes with number of elements ranging from 30,269 elements for the Barth5 mesh to 11,451 elements for the Barth4 mesh. Future work with PART entails using the tool with an integrated application requiring distributed systems. In particular this application, illustrated in the document entails an integration of finite element and fluid dynamic simulations to address the cooling of turbine blades of a gas turbine engine design. It is not uncommon to encounter high-temperature, film-cooled turbine airfoils with 1,000,000s of degrees of freedom. This results because of the complexity of the various components of the airfoils, requiring fine-grain meshing for accuracy. Additional information is contained in the original.

Taylor, Valerie↗