Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Benchmarking Software”

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 325 records · Page 18
All sourcesNASA NTRSDOE OSTI
DOE OSTI2020

FY20 Proxy App Suite Release: Report for ECP Proxy App Project Milestone ADCD-504-10

Version 4.0 of the ECP Proxy App Suite is practically unchanged from the previous release. The current set of proxies has proven useful for many aspects of benchmarking and co-design and we see little reason to alter the suite. Although there have been few changes to the ECP suite, the team has been hard at work in other areas. In the area of Machine Learning (ML) we have now created a separate proxy suite dedicated to this scientific applications of ML. The suite includes: miniGAN (Generative Adversarial Networks), miniRL (Reinforcement Learning), CRADL (inline inference), Cosmoflow-Benchmark (Convolutional Neural Network), and MLPerf-DeepCam (Climate Segmentation Benchmark). Section 3 contains more information about these proxies as well as the principles that are guiding the development of the suite. We have surveyed available proxies for several application domains including Computational Fluid Dynamics, Quantum Chemistry, Quantum Computing Simulation, Molecular Dynamics, Monte Carlo Transport, and Density Functional Theory to identify gaps in proxy coverage. Several new proxy apps are either already available or will be released soon to fill these gaps. Section 4 provides full details. Finally, section 5 reports on our continued collaboration with the ECP Continuous Integration (CI) effort to use proxy apps to help identify problems and roadblocks to cross-lab CI. We have also assisted the El Capitan Center of Excellence (COE) to stand up a CI system that will be used to test releases of HPE and AMD software stacks. We hope that the COE effort can serve as a model for ECP by showing how the Proxy App Team can work with AD and ST teams identify critical features, kernels, patterns, etc. and incorporate them into a CI system that will help ensure that Frontier and Aurora will provide those needed capabilities.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2026

V-HAMSTeR v1.0.0

V-HAMSTeR is a bioinformatics software tool designed to predict the hosts of viruses directly from genomic sequences. It can be used by researchers to predict animal, prokaryotic, plant, protist or fungal viral hosts including viruses that may be fragmented or discovered in environmental metagenomic datasets. Features & Uses: The software employs a novel dual-stream deep learning architecture that dynamically fuses implicit sequence embeddings from a genomic foundation model with 13 explicit, handcrafted biological features (e.g., coding density and strand switch rates). To ensure maximum reliability, V=HAMSTeR deploys a 5-fold deep ensemble calibrated via Joint Temperature Scaling, providing users with statistically rigorous confidence probabilities. It also features an automated sequence chunking and mean-pooling module to seamlessly process variable-length contigs. Advantages Over Similar Technologies: Existing tools (e.g., IPEV, RNAVirHost) typically rely on either basic k-mers or isolated neural networks. V-HAMSTeR's hybrid architecture captures both broad genomic context and specific biological motifs that standalone foundation models often miss. Furthermore, unlike competitor tools that struggle with incomplete data or exhibit extreme overconfidence, V-HAMSTeR is explicitly benchmarked and mathematically calibrated for fragmented assemblies (1kb–10kb). This makes it uniquely robust, accurate, and trustworthy for the messy reality of real-world environmental viromics.

Grigson, Susie [Lawrence Berkeley National Laborat↗
DOE OSTI2021

Radiological Anomaly Detection And Identification (RADAI) v1.0

The Radiological Anomaly Detection and Identification (RADAI) software package is a python library for implementing, training, and storing algorithms that detect and identify anomalies in gamma-ray spectra. The library defines a general framework for implementing detection (binary) and identification (classification) algorithms, objects to encapsulate the results of analyses, a variety of temporal filtering tools that can be used in constructing algorithms, and conceptual design that allows easy reading and writing of algorithms (and their time dependent state). In addition to this framework, the library includes implementations of a variety of algorithms from the scientific literature including: gross-counts k-sigma, SPRT, N-SCRAD, Region of Interest, and Censored Energy Window. The implementation of these algorithms within the RADAI package was done to facilitate user-initiated training and configuration to by applied to different gamma-ray detector types. Finally, benchmarked and synthetic datasets will be made available for standardized algorithm characterization with corresponding utilities in the RADAI package for data access and processing.

Joshi, Tenzing↗
DOE OSTI2021

Machine Learning-enabled Scalable Performance Prediction of Scientific Codes

Hardware architectures become increasingly complex as the compute capabilities grow to exascale. Here, we present the Analytical Memory Model with Pipelines (AMMP) of the Performance Prediction Toolkit (PPT). PPT-AMMP takes high-level source code and hardware architecture parameters as input and predicts runtime of that code on the target hardware platform, which is defined in the input parameters. PPT-AMMP transforms the code to an (architecture-independent) intermediate representation, then (i) analyzes the basic block structure of the code, (ii) processes architecture-independent virtual memory access patterns that it uses to build memory reuse distance distribution models for each basic block, and (iii) runs detailed basic-block level simulations to determine hardware pipeline usage. PPT-AMMP uses machine learning and regression techniques to build the prediction models based on small instances of the input code, then integrates into a higher-order discrete-event simulation model of PPT running on Simian PDES engine. We validate PPT-AMMP on four standard computational physics benchmarks and present a use case of hardware parameter sensitivity analysis to identify bottleneck hardware resources on different code inputs. We further extend PPT-AMMP to predict the performance of a scientific application code, namely, the radiation transport mini-app SNAP. To this end, we analyze multi-variate regression models that accurately predict the reuse profiles and the basic block counts. We validate predicted SNAP runtimes against actual measured times.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2022

Code Verification and Solution Verification framework in pin-resolved neutron transport code MPACT

Program verification in scientific computing encompasses the application of formal and mathematical techniques to a scientific computing code for its credibility, accuracy, and validity. Code Verification identifies bugs and performance issues in the software development stage. Solution Verification assesses the applicability of the code and the accuracy of the solution to problems of interest. Both activities utilize application cases and quantify the error against prescribed acceptance criteria. However, simply executing more application cases does not guarantee stronger or more comprehensive credibility. Here, we establish a verification framework that involves Code Verification and Solution Verification, both of which work together such that the overarching goal of “converge to the correct answer for the intended application” can be reasonably inferred. The application of such a verification framework is demonstrated using the pin-resolved neutron transport code MPACT, where standard unit tests and regression tests are covered, and where the Method of Exact Solutions and the Method of Manufactured Solutions are successfully used. Additionally, the applicability of Method of Manufactured Solutions is extended to the OECD/NEA C5G7 benchmark problems of practical material and geometric configurations. Solution Verification activities are demonstrated on a practical hierarchy of application models of increasing complexity ranging from 2D pin cell problems to 3D assembly problems. The convergence behavior and rate of convergence with respect to each individual variable are studied and provided. This framework can be adapted broadly to other fields involving scientific computing codes.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2023

Identification of mobile genetic elements with geNomad

Identifying and characterizing mobile genetic elements in sequencing data is essential for understanding their diversity, ecology, biotechnological applications and impact on public health. Here we introduce geNomad, a classification and annotation framework that combines information from gene content and a deep neural network to identify sequences of plasmids and viruses. geNomad uses a dataset of more than 200,000 marker protein profiles to provide functional gene annotation and taxonomic assignment of viral genomes. Using a conditional random field model, geNomad also detects proviruses integrated into host genomes with high precision. In benchmarks, geNomad achieved high classification performance for diverse plasmids and viruses (Matthews correlation coefficient of 77.8% and 95.3%, respectively), substantially outperforming other tools. Leveraging geNomad’s speed and scalability, we processed over 2.7 trillion base pairs of sequencing data, leading to the discovery of millions of viruses and plasmids that are available through the IMG/VR and IMG/PR databases. geNomad is available at https://portal.nersc.gov/genomad.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2021

A Poisson equation method for prescribing fully developed non-Newtonian inlet conditions for computational fluid dynamics simulations in models of arbitrary cross-section

Prescribing inlet boundary conditions for computational fluid dynamics (CFD) simulations of internal flow in complex geometries such as anatomical vascular models is challenging. In the absence of patient-specific inlet velocity data, a common approach for long blood vessels is to assume that the inlet flow is fully developed. In vessels of irregular cross section, however, prescribing fully developed conditions is complicated due to the lack of a general closed-form analytical solution. In this study, we develop a simple Poisson equation method for prescribing fully developed inlet conditions for the flow of either Newtonian or non-Newtonian fluids in CFD models of arbitrary cross-section. We first derive the generalized Poisson equation for fully developed flow of a non-Newtonian fluid and we then develop and verify a methodology for numerically computing the solution on any planar boundary domain. In addition, we develop a simple extension of the method for prescribing a non-orthogonal inlet velocity that represents fully developed flow from an upstream tube that is connected to the CFD inlet at a non-orthogonal angle. This may be used to investigate a common source of uncertainty in CFD simulations of internal flow that is due to a lack of information concerning the exact streamwise flow direction at the inlets. Comparison to several Newtonian and non-Newtonian benchmark verification solutions shows the method to be extremely accurate. As a practical demonstration case, we use the method to prescribe fully developed conditions on multiple non-circular inlets for the non-Newtonian flow of blood in a patient-specific model of the inferior vena cava (IVC). Finally, we further demonstrate the utility of the method by performing a sensitivity study using the patient-specific IVC model, wherein we investigate the influence of inlet velocity flow direction on the non-Newtonian IVC hemodynamics. Given its simplicity and computational efficiency, the method is shown to be far superior to alternative approaches for prescribing fully developed inlet conditions in such complicated geometries. In conclusion, to facilitate the adoption of our Poisson equation method, we have distributed our OpenFOAM source code and the associated test cases from this study as open-source software.

97 MATHEMATICS AND COMPUTING↗
DOE OSTI2025

Three-Dimensional Heat Flux and Thermal Analysis of Angled Tungsten Samples on DIII-D

ITER-grade tungsten and dispersoid-strengthened tungsten samples with the top surface angled at ~15° towards the incident plasma flux were exposed to 9 H-mode discharges with edge-localized modes (ELMs) in the lower divertor of DIII-D tokamak using the Divertor Material Evaluation System (DiMES). Surface damage included cracking and flaking of material on the two samples farthest away from the plasma strike point, and significant melting of the two samples closest to the strike point. Heat flux and thermal analysis tools new to DIII-D have been applied to better understand this material response and to help optimize the exposure conditions for future experiments. SMITER field-line tracing simulations based on IRTV data and EFIT equilibria estimate an average inter-ELM perpendicular heat flux, 𝑞⊥,𝑖nter−𝐸LM , on the angled surfaces of 10.1 – 19.6 MW/m² for a majority of the 9 discharges, increasing to 15.6 – 24.5 MW/m² for the single, higher-power shot where samples melted. Fast camera data showed shallow intra-ELM melting and re-solidification, which transitioned to bulk inter-ELM melting with melt motion in the 𝐽⃗ 𝑥 𝐵⃗ direction. About 50% of the protruding volume of the most affected sample was displaced via melt-motion. SIERRA thermal modeling software was able to reproduce an onset time of melting consistent with fast camera data and final sample conditions, within < 200 ms. Maximum surface temperatures of 3122 K and 2787 K are estimated for the samples farthest away from the strike point, while the closest samples achieve melting at 4067 ms and 4750 ms into the ~5000 ms plasma exposure. A +10% increase in both the SMITER 𝑞⊥,𝑖nter−𝐸LM calculations and the estimated ELM heat loads 𝑞⊥, 𝐸LM was required to achieve this result, which is within the uncertainty of the diagnostic data but likely accounts for non-ideal geometry effects plus other physics uncertainties not included in this first iteration of modeling. This work provided valuable estimates of the 3D temperature evolution to help better understand the observed surface morphology and internal recrystallization of samples, which are discussed in detail in a complementary manuscript [1]. Benchmarking efforts with more diagnosed DIII-D experiments are underway to further refine the SMITER and SIERRA models for DiMES. Future use of these tools will enable researchers to precisely target heat flux exposure conditions in DIII-D to test, but not exceed, the thermomechanical limitations of novel plasma-facing materials.

70 PLASMA PHYSICS AND FUSION TECHNOLOGY↗
DOE OSTI2026

Knowledge Graph for End-to-End Traceability of an Integrated Human-Earth System Model

Integrated human-Earth system models inform energy-water-land system dynamics and policies, yet their results are difficult to trace through input-data, model structure, scenario configurations, and solved outputs. Because this information is siloed across disconnected artifacts, process-based IAMs have historically lacked a unified, queryable representation. Such lack of traceability prevents researchers from systematically isolating the multi-sector drivers of complex outcomes (such as tracing water-scarcity results back to distant energy-system dynamics) or conducting holistic uncertainty attribution across hundreds of interacting parameters. To address this concern, our work documents the software engineering process of a knowledge graph that unifies these four layers for the Global Change Analysis Model (GCAM-USA_Reference scenario, GCAM v9.1). The graph was built as a relational property graph in DuckDB from the run’s own artifacts: the input-preparation dependency map (gcamdata chunk map), the model’s XML input files, the run configuration, and the results database (BaseX), successfully mapping the model’s declared structure. The resulting graph comprises 204,321 nodes and 1,687,814 edges across 16 node types and 15 edge types, with approximately 16.3 million time-series values stored separately to maintain structural efficiency. To ensure representation fidelity, every edge carries an epistemic-status annotation recording the warrant for the relationship (structural, provenance, dependency, or model-derived), and a machine-readable provenance ledger classifying the origin of every schema element. Evaluation against a fixed five-benchmark suite with locked baselines reports zero structural orphans, zero dangling edge endpoints, and 100% of output-producing technologies traceable to raw input files. Two interactive interfaces present the graph, including a serverless browser application built on DuckDB-Wasm. By establishing the first end-to-end provenance framework for an IAM, this work enables researchers and scientists to systematically audit complex policy scenarios, debug model structures, and trace policy-relevant outputs to their data origins in real time.

Artifical Intelligence↗
DOE OSTI2020

JETSCAPE framework: p+p results

The JETSCAPE framework is a modular and versatile Monte Carlo software package for the simulation of high energy nuclear collisions. In this work we present a new tune of JETSCAPE, called PP19, and validate it by comparison to jet-based measurements in p + p collisions, including inclusive single jet cross sections, jet shape observables, fragmentation functions, charged hadron cross sections, and dijet mass cross sections. These observables in p+p collisions provide the baseline for their counterparts in nuclear collisions. Quantifying the level of agreement of JETSCAPE results with p+p data is thus necessary for meaningful applications of JETSCAPE to A+A collisions. The calculations use the JETSCAPE PP19 tune, defined in this paper, based on version 1.0 of the JETSCAPE framework. For the observables discussed in this work calculations using JETSCAPE PP19 agree with data over a wide range of collision energies at a level comparable to standard Monte Carlo codes. These results demonstrate the physics capabilities of the JETSCAPE framework and provide benchmarks for JETSCAPE users.

72 PHYSICS OF ELEMENTARY PARTICLES AND FIELDS↗
DOE OSTI2022

Sensitivity Coefficients Calculated for the Prompt Neutron Decay Constant at or Near Delayed Critical [Abstract]

Experimenters at Los Alamos National Laboratory (LANL) measure the prompt neutron decay constant for many experiments at the National Criticality Experiments Research Center (NCERC) to infer reactivity and the effective neutron multiplication factor. These quantities are very important for nuclear criticality safety and validating nuclear data. Uncertainty in measures of criticality of an experimental configuration can be determined prior to physically performing the experiment by applying first order perturbation theory to Monte Carlo codes, such as MCNP®. The first order perturbation theory produces first derivatives of some nuclear parameter to nuclear data (e.g., cross section data). This first derivative is commonly referred to as a sensitivity coefficient. Currently, the MCNP® software has the capability of computing effective neutron multiplication factor sensitivity coefficients to cross section data. This work builds off of this MCNP® capability and the first order perturbation theory to provide a method of calculating sensitivity coefficients for the prompt neutron decay constant at or near delayed critical to cross section data. The prompt neutron decay constant sensitivity coefficient calculated in this work does not depend on any modification of the MCNP® source code. Prompt neutron decay constant sensitivity coefficient calculations can be used to infer reactivity and effective neutron multiplication factor sensitivity coefficient values as well. By investigating the trends of prompt neutron decay constant sensitivity coefficients for nuclide-reaction pairs across energy spectra, experiments can be designed to maximize or minimize the uncertainty in the prompt neutron decay constant in a particular energy region, which can lead to further optimization studies. Prompt neutron decay constant sensitivity coefficients will be calculated for the Jezebel benchmark. Subsequently, these sensitivity coefficients will be used in a data assimilation process to determine if there is or are optimal experiments that can be performed to provide insight into adjustments of uncertain/inaccurate cross section data. Specifically, the effect of the prompt neutron decay constant sensitivity coefficients on the nuclear data-induced uncertainty in the effective neutron multiplication factor will be examined

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗
DOE OSTI2024

Community Geothermal: Energy, Cost, and Carbon Modeling for District Design - Ann Arbor, MI

This data includes results on an analysis of existing and projected energy, cost, and carbon for the City of Ann Arbor - District Geothermal Design and Deployment to Equitably Decarbonize Low Income Neighborhoods in Ann Arbor project. The scope of the project includes designing and implementing a geothermal district heating and cooling system that reduces thermal heating and cooling load by 75% and greenhouse gas emissions by 40% in the project area (262 households, 6 commercial buildings). The existing neighborhood was modeled using Design Builder, an EnergyPlus software, to understand the current energy load. The energy model was then flipped to reflect the designed district geothermal heating and cooling system to project the effect on energy, carbon, and cost. This dataset includes the analysis files utilized and created for this study. There are 3 categories of data: 1) existing/benchmarking, 2) energy modeling, and 3) post processed calculations. This follows the methodology and process of the project team, which is fully explained in file 00_Technical Economic Environmental Assessment. All uses of data are referenced throughout this assessment to their respective files included below.

15 GEOTHERMAL ENERGY↗
DOE OSTI2023

Application Experiences on a GPU-Accelerated Arm-based HPC Testbed

This paper assesses and reports the experience of ten teams working to port, validate, and benchmark several High Performance Computing applications on a novel GPU-accelerated Arm testbed system. The testbed consists of eight NVIDIA Arm HPC Developer Kit systems, each one equipped with a server-class Arm CPU from Ampere Computing and two data center GPUs from NVIDIA Corp. The systems are connected together using InfiniBand interconnect. The selected applications and mini-apps are written using several programming languages and use multiple accelerator-based programming models for GPUs such as CUDA, OpenACC, and OpenMP offloading. Working on application porting requires a robust and easy-to-access programming environment, including a variety of compilers and optimized scientific libraries. The goal of this work is to evaluate platform readiness and assess the effort required from developers to deploy well-established scientific workloads on current and future generation Arm-based GPU-accelerated HPC systems. The reported case studies demonstrate that the current level of maturity and diversity of software and tools is already adequate for large-scale production deployments.

Elwasif, Wael↗
DOE OSTI2022

sciCAN: single-cell chromatin accessibility and gene expression data integration via cycle-consistent adversarial network

The boom in single-cell technologies has brought a surge of high dimensional data that come from different sources and represent cellular systems from different views. With advances in these single-cell technologies, integrating single-cell data across modalities arises as a new computational challenge. Here, we present an adversarial approach, sciCAN, to integrate single-cell chromatin accessibility and gene expression data in an unsupervised manner. We benchmarked sciCAN with 5 existing methods in 5 scATAC-seq/scRNA-seq datasets, and we demonstrated that our method dealt with data integration with consistent performance across datasets and better balance of mutual transferring between modalities than the other 5 existing methods. We further applied sciCAN to 10X Multiome data and confirmed that the integrated representation preserves biological relationships within the hematopoietic hierarchy. Finally, we investigated CRISPR-perturbed single-cell K562 ATAC-seq and RNA-seq data to identify cells with related responses to different perturbations in these different modalities.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2020

Improving the performance of medical imaging applications using SYCL

As opposed to the Open Computing Language (OpenCL) programming model, in which host and device codes are generally written in different programming languages, SYCL can combine host and device codes for an application in a type-safe way to improve development productivity and performance portability. Hence, this report shows the experimental results of applying the SYCL programming model to medical imaging applications for a study on performance portability and programming productivity between OpenCL and SYCL. Rodinia is a widely used open-source benchmark suite for heterogeneous computing. We choose two medical imaging applications (Heart Wall and Particle Filter) in the benchmark suite, migrate the OpenCL implementations of the applications to the SYCL implementations, and evaluate their performance and productivity on Intel® microprocessors that contain a central processing unit (CPU) and an integrated graphics processing unit (GPU). The maturing SYCL compilers, which are based on a conformant implementation of the SYCL 1.2.1 Khronos specification, have been actively optimized for Intel® computing platforms. The experimental results are promising in terms of the raw performance and productivity. Although the SYCL implementation of the Heart Wall application does not execute successfully on a CPU, the SYCL implementations of the application can achieve comparable or better performance than the OpenCL implementation on Intel® integrated GPUs. For the Particle Filter application, the performance difference between the SYCL and OpenCL implementations is comparable on the GPUs for most cases, but the SYCL implementations are on average 4X faster than the OpenCL implementations on Intel® Xeon® CPUs. For programming productivity, we arguably use lines of code as a way to measure programming productivity in software. The SYCL programs reduce the lines of code of the OpenCL programs by 52% and 38% for the Heart Wall and Particle Filter, respectively. The results indicate that SYCL is a promising programming model for heterogeneous computing with the maturing compilers. We organize the remainder of the report as follows. Section II introduces the SYCL programming model, compares the major differences between an OpenCL program and a SYCL program, and describes the characteristics of the two applications. Section III describes the SYCL programming model in more details, and shows the SYCL implementation of a kernel function in the Particle Filter application as an example. In Section IV, we evaluate the performance of the applications on the CPUs and GPUs. Section V concludes the report.

59 BASIC BIOLOGICAL SCIENCES↗
DOE OSTI2022

The EBLM project – IX. Five fully convective M-dwarfs, precisely measured with CHEOPS and TESS light curves

ABSTRACT Eclipsing binaries are important benchmark objects to test and calibrate stellar structure and evolution models. This is especially true for binaries with a fully convective M-dwarf component for which direct measurements of these stars’ masses and radii are difficult using other techniques. Within the potential of M-dwarfs to be exoplanet host stars, the accuracy of theoretical predictions of their radius and effective temperature as a function of their mass is an active topic of discussion. Not only the parameters of transiting exoplanets but also the success of future atmospheric characterization relies on accurate theoretical predictions. We present the analysis of five eclipsing binaries with low-mass stellar companions out of a subsample of 23, for which we obtained ultra-high-precision light curves using the CHEOPS satellite. The observation of their primary and secondary eclipses are combined with spectroscopic measurements to precisely model the primary parameters and derive the M-dwarfs mass, radius, surface gravity, and effective temperature estimates using the PYCHEOPS data analysis software. Combining these results to the same set of parameters derived from TESS light curves, we find very good agreement (better than 1 per cent for radius and better than 0.2 per cent for surface gravity). We also analyse the importance of precise orbits from radial velocity measurements and find them to be crucial to derive M-dwarf radii in a regime below 5 per cent accuracy. These results add five valuable data points to the mass–radius diagram of fully convective M-dwarfs.

Sebastian, D. (ORCID:0000000222149258)↗
DOE OSTI2020

Hindsight logging for model training

In modern Machine Learning, model training is an iterative, experimental process that can consume enormous computation resources and developer time. To aid in that process, experienced model developers log and visualize program variables during training runs. Exhaustive logging of all variables is infeasible, so developers are left to choose between slowing down training via extensive conservative logging, or letting training run fast via minimalist optimistic logging that may omit key information. As a compromise, optimistic logging can be accompanied by program checkpoints; this allows developers to add log statements post-hoc, and "replay" desired log statements from checkpoint---a process we refer to as hindsight logging. Unfortunately, hindsight logging raises tricky problems in data management and software engineering. Done poorly, hindsight logging can waste resources and generate technical debt embodied in multiple variants of training code. In this paper, we present methodologies for efficient and effective logging practices for model training, with a focus on techniques for hindsight logging. Our goal is for experienced model developers to learn and adopt these practices. To make this easier, we provide an open-source suite of tools for Fast Low-Overhead Recovery (flor) that embodies our design across three tasks: (i) efficient background logging in Python, (ii) adaptive periodic checkpointing, and (iii) an instrumentation library that codifies hindsight logging for efficient and automatic record-replay of model-training. Model developers can use each flor tool separately as they see fit, or they can use flor in hands-free mode, entrusting it to instrument their code end-to-end for efficient record-replay. Our solutions leverage techniques from physiological transaction logs and recovery in database systems. Evaluations on modern ML benchmarks demonstrate that flor can produce fast checkpointing with small user-specifiable overheads (e.g. 7%), and still provide hindsight log replay times orders of magnitude faster than restarting training from scratch.

Computer Science↗
DOE OSTI2023

HIPLZ: Enabling performance portability for exascale systems

While heterogeneous computing has emerged as a dominant trend in current and future High-Performance Computing (HPC) systems, it is also widely recognized that this shift has led to increased software complexity due to a proliferation of programming systems for different heterogeneous processors. One such example is the Heterogeneous-Compute Interface for Portability from AMD (HIP ), which is composed of a C Runtime API and C++ Kernel Language. Many HPC applications will likely use HIP on future exascale systems (e.g., Frontier and El Capitan), but HIP currently only targets AMD and NVIDIA processors. This limitation creates challenges for users who would also like to run their applications on exascale systems based on other architectures (e.g., Aurora, which is based on Intel hardware) that are currently not targeted by HIP . In this paper, we introduce the design and implementation of HIPLZ , a compiler and runtime system that uses the Intel Level Zero API to support HIP on Intel GPU architectures. We discuss the design of HIPLZ , derived from HIPCL (an implementation of HIP on top of OpenCL ), and portability issues that occur from using the Level Zero runtime as a backend. We evaluate our implementation by running several performance benchmarks and mini-apps written in HIP on Intel architectures using HIPLZ . Our results show that this approach provides competitive performance relative to Intel's OpenCL implementations on Intel Gen9 and UHD Graphics 770 GPUs, while providing good coverage of features needed by HPC applications. Overall, this approach is a promising demonstration of enabling performance portability for exascale systems.

97 MATHEMATICS AND COMPUTING↗