Umpire: Application-focused management and coordination of complex hierarchical memory

Beckingsale, David; McFadden, Marty; Dahm, Johann; Pankajakshan, Ramesh; Hornung, Richard D.

doi:10.1147/jrd.2019.2954403

Cited by 30 publications

(19 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For many large scale LLNL simulation codes, running efficiently on Sierra required major code refactoring such as porting computational kernels and revisiting memory management strategies. Abstraction layers such as RAJA [9] and memory resource managers such as Umpire [10] were utilized to simplify the porting process while enabling a single source code for different computing platforms. Expertise from developers and domain scientists was required to ensure correctness and performance.…”

Section: Overview Of Tested Simulation Codesmentioning

confidence: 99%

“…Some of its key capabilities include multi-material radiation hydrodynamics in both an Eulerian and an Arbitrary Lagrangian-Eulerian (ALE) framework for applications in inertial confinement fusion (ICF), pulsed power and equation of state/material strength experiments. MARBL builds on modular physics and computer science packages such as Axom [14], MFEM [15], RAJA [9], and Umpire [10]. A distinct feature of this code is its design choice of employing high-order numerical methods.…”

Section: Marblmentioning

confidence: 99%

See 1 more Smart Citation

Understanding Power and Energy Utilization in Large Scale Production Physics Simulation Codes

Ryujin¹,

Vargas²,

Karlin³

et al. 2022

Preprint

View full text Add to dashboard Cite

Power is an often-cited reason for moving to advanced architectures on the path to Exascale computing. This is due to the practical concern of delivering enough power to successfully site and operate these machines, as well as concerns over energy usage while running large simulations. Since accurate power measurements can be difficult to obtain, processor thermal design power (TDP) is a possible surrogate due to its simplicity and availability. However, TDP is not indicative of typical power usage while running simulations. Using commodity and advance technology systems at Lawrence Livermore National Laboratory (LLNL) and Sandia National Laboratory, we performed a series of experiments to measure power and energy usage in running simulation codes. These experiments indicate that large scale LLNL simulation codes are significantly more efficient than a simple processor TDP model might suggest.

show abstract

Section: Overview Of Tested Simulation Codesmentioning

confidence: 99%

Section: Marblmentioning

confidence: 99%

Understanding Power and Energy Utilization in Large Scale Production Physics Simulation Codes

Ryujin¹,

Vargas²,

Karlin³

et al. 2022

Preprint

View full text Add to dashboard Cite

show abstract

“…ExaSGD adopted a programming model based on C++ and portability libraries RAJA [43] and Umpire [44] for key software stack components-HiOp and ExaGO. Using portability libraries greatly increased the speed of ExaSGD software development and reduced the development cost.…”

Section: Programming Modelmentioning

confidence: 99%

Map Applications to Target Exascale Architecture with Machine-Specific Performance Analysis, Including Challenges and Projections

Siegel

Evans

Draeger

et al. 2021

View full text Add to dashboard Cite

show abstract

“…To handle these memory management complexities, many applications provide memory management data structures, or employ a memory management library such as Umpire [10]. To support these applications, SUNDIALS provides a set of utilities collectively referred to as the SUNMemoryHelper API.…”

Section: Memory Managementmentioning

confidence: 99%

Enabling GPU Accelerated Computing in the SUNDIALS Time Integration Library

Balos,

Gardner,

Woodward

et al. 2020

Preprint

View full text Add to dashboard Cite

As part of the Exascale Computing Project (ECP), a recent focus of development efforts for the SUite of Nonlinear and DIfferential/ALgebraic equation Solvers (SUNDIALS) has been to enable GPU-accelerated time integration in scientific applications at extreme scales. This effort has resulted in several new GPU-enabled implementations of core SUNDIALS data structures, support for programming paradigms which are aware of the heterogeneous architectures, and the introduction of utilities to provide new points of flexibility. In this paper, we discuss our considerations, both internal and external, when designing these new features and present the features themselves. We also present performance results for several of the features on the Summit supercomputer and early access hardware for the Frontier supercomputer, which demonstrate negligible performance overhead resulting from the additional infrastructure and significant speedups when using both NVIDIA and AMD GPUs.

show abstract

Umpire: Application-focused management and coordination of complex hierarchical memory

Cited by 30 publications

References 5 publications

Understanding Power and Energy Utilization in Large Scale Production Physics Simulation Codes

Understanding Power and Energy Utilization in Large Scale Production Physics Simulation Codes

Map Applications to Target Exascale Architecture with Machine-Specific Performance Analysis, Including Challenges and Projections

Enabling GPU Accelerated Computing in the SUNDIALS Time Integration Library

Contact Info

Product

Resources

About