SEARCH · Search NASA

Results for “RMA”

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.

SourceResearch connectionFrom yearThrough year

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

At least 37 records · Page 2

All sources NASA NTRS DOE OSTI

DOE OSTI2015

Phase Diagram of a Three-Dimensional Antiferromagnet with Random Magnetic Anisotropy

Three-dimensional (3D) antiferromagnets with random magnetic anisotropy (RMA) that were experimentally studied to date have competing two-dimensional and three-dimensional exchange interactions which can obscure the authentic effects of RMA. The magnetic phase diagram of Fe x Ni 1-x F 2 epitaxial thin films with true random single-ion anisotropy was deduced from magnetometry and neutron scattering measurements and analyzed using mean field theory. Regions with uniaxial, oblique and easy plane anisotropies were identified. A RMA-induced glass region was discovered where a Griffiths-like breakdown of long-range spin order occurs.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2019

Exploring MPI Communication Models for Graph Applications Using Graph Matching as a Case Study

Traditional implementations of parallel graph operations on distributed memory platforms are written using Message Passing Interface (MPI) point to point communication primitives such as Send-Recv (blocking and nonblocking). Apart from this classical model, the MPI community has over the years added other communication models; however, their suitability for handling the irregular traffic workloads typical of graph operations remain comparatively less explored. Our aim in this paper is to study these relatively underutilized communication models of MPI for graph applications. More specifically, we evaluate MPI’s one-sided programming, or Remote Memory Access (RMA), and nearest neighborhood collectives using a process graph topology. There are features in these newer models that are intended to better map to irregular communication patterns, as exemplified in graph algorithms. As a concrete application for our case study, we use distributed memory implementations of a weighted graph matching algorithm to investigate performances of MPI-3 RMA and neighborhood collective operations compared to nonblocking Send- Recv. A matching in a graph is a subset of edges such that no two matched edges are incident on the same vertex. A maximum weight matching is a matching of maximum weight computed as the sum of the weights of matched edges. Algorithms for optimal matching have quadratic complexity and are difficult to parallelize. More importantly, execution of graph matching is dominated by high volume of irregular memory accesses, making it an ideal candidate for studying the effects of various MPI communication models on graph applications at scale. Our neighborhood collectives and RMA implementations yields up to 6× speedup over traditional optimized nonblocking Send-Recv implementations, on thousands of cores of the NERSC Cori supercomputer. We believe the lessons learned from this case study can be adopted to benefit a wider range of graph application codes.

Ghosh, Sayan↗

DOE OSTI2015

SU-F-BRD-01: A Novel 4D Robust Optimization Mitigates Interplay Effect in Intensity-Modulated Proton Therapy for Lung Cancer

Purpose: To compare the impact of interplay effect on 3D and 4D robustly optimized intensity-modulated proton therapy (IMPT) plans to treat lung cancer. Methods: Two IMPT plans were created for 11 non-small-cell-lung-cancer cases with 6–14 mm spots. 3D robust optimization generated plans on average CTs with the internal gross tumor volume density overridden to deliver 66 CGyE in 33 fractions to the internal target volume (ITV). 4D robust optimization generated plans on 4D CTs with the delivery of prescribed dose to the clinical target volume (CTV). In 4D optimization, the CTV of individual 4D CT phases received non-uniform doses to achieve a uniform cumulative dose. Dose evaluation software was developed to model time-dependent spot delivery to incorporate interplay effect with randomized starting phases of each field per fraction. Patient anatomy voxels were mapped from phase to phase via deformable image registration to score doses. Indices from dose-volume histograms were used to compare target coverage, dose homogeneity, and normal-tissue sparing. DVH indices were compared using Wilcoxon test. Results: Given the presence of interplay effect, 4D robust optimization produced IMPT plans with better target coverage and homogeneity, but slightly worse normal tissue sparing compared to 3D robust optimization (unit: Gy) [D95% ITV: 63.5 vs 62.0 (p=0.014), D5% - D95% ITV: 6.2 vs 7.3 (p=0.37), D1% spinal cord: 29.0 vs 29.5 (p=0.52), Dmean total lung: 14.8 vs 14.5 (p=0.12), D33% esophagus: 33.6 vs 33.1 (p=0.28)]. The improvement of target coverage (D95%,4D – D95%,3D) was related to the ratio RMA3/(TVx10−4), with RMA and TV being respiratory motion amplitude (RMA) and tumor volume (TV), respectively. Peak benefit was observed at ratios between 2 and 10. This corresponds to 125 – 625 cm3 TV with 0.5-cm RMA. Conclusion: 4D optimization produced more interplay-effect-resistant plans compared to 3D optimization. It is most effective when respiratory motion is modest compared to TV. NIH/NCI K25CA168984; Eagles Cancer Research Career Development; The Lawrence W. and Marilyn W. Matteson Fund for Cancer Research; Mayo ASU Seed Grant; The Kemper Marley Foundation.

60 APPLIED LIFE SCIENCES↗

DOE OSTI2017

Merged Requests for Better Performance and Productivity in Multithreaded OpenSHMEM

A merged request is a handle representing a group of Remote Memory Access (RMA), Atomic or Collective operations. The merged request can be created either by combining multiple outstanding merged request handles or using the same merged request handle for additional operations. We show that introducing such simple yet powerful semantics in OpenSHMEM provides many productivity and performance advantages. In this paper, we first introduce the interfaces and semantics for creating and using merged request handles. Then, we demonstrate with a merge request that we can achieve better performance characteristics in multithreaded OpenSHMEM application. Particularly, we show one can achieve higher message rate, a higher bandwidth for smaller message, and better computation-communication overlap. Further, we use merged request to realize multithreaded collectives, where multiple threads co-operate to complete the collective operation. Our experimental results show that in a multithreaded OpenSHMEM program, the merged request based RMA operations achieve over 100 Million Messages Per Second (MMPS). It achieves over 10 MMPS compared to 4.5 MMPS with default RMA operations in a single threaded environment. Also, we achieve higher bandwidth for smaller message sizes, close to 100% overlap, and reduce the latency by 60%.

Gorentla Venkata, Manjunath↗

DOE OSTI2024

Effective R-Matrix Parameterizations for Nuclear Data

Explore the source record for details and available documents.

Arbanas, Goran↗

DOE OSTI2016

OpenMPI and ExxonMobil Topics

These are a series of slides for a presentation for ExxonMobil's visit to Los Alamos National Laboratory. Topics covered are: Open MPI - The Release Story, MPI-3 RMA in Open MPI, MPI dynamic process management and Open MPI, and new options with CLE 6. Open MPI RMA features are: since v2.0.0 full support for the MPI-3.1 specification, support for non-contiguous datatypes, support for direct use of the RDMA capabilities of high performance networks (Cray Gemini/Aries, Infiniband), starting in v2.1.0 will have support for using network atomic operations for MPI_Fetch_and_op and MPI_Compare_and_swap, tested with MPI_THREAD_MULTIPLE.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2014

Rovibrational energy transfer in the He-C{sub 3} collision: Potential energy surface and bound states

We present a four-dimensional potential energy surface (PES) for the collision of C{sub 3} with He. Ab initio calculations were carried out at the coupled-cluster level with single and double excitations and a perturbative treatment of triple excitations, using a quadruple-zeta basis set and mid-bond functions. The global minimum of the potential energy is found to be −26.9 cm{sup −1} and corresponds to an almost T-shaped structure of the van der Waals complex along with a slightly bent configuration of C{sub 3}. This PES is used to determine the rovibrational energy levels of the He-C{sub 3} complex using the rigid monomer approximation (RMA) and the recently developed atom-rigid bender approach at the Close Coupling level (RB-CC). The calculated dissociation energies are −9.56 cm{sup −1} and −9.73 cm{sup −1}, respectively at the RMA and RB-CC levels. This is the first theoretical prediction of the bound levels of the He-C{sub 3} complex with the bending motion.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗

DOE OSTI2014

On the Suitability of MPI as a PGAS Runtime

Partitioned Global Address Space (PGAS) models are emerging as a popular alternative to MPI models for designing scalable applications. At the same time, MPI remains a ubiquitous communication subsystem due to its standardization, high performance, and availability on leading platforms. In this paper, we explore the suitability of using MPI as a scalable PGAS communication subsystem. We focus on the Remote Memory Access (RMA) communication in PGAS models which typically includes {\em get, put,} and {\em atomic memory operations}. We perform an in-depth exploration of design alternatives based on MPI. These alternatives include using a semantically-matching interface such as MPI-RMA, as well as not-so-intuitive interfaces such as MPI two-sided with a combination of multi-threading and dynamic process management. With an in-depth exploration of these alternatives and their shortcomings, we propose a novel design which is facilitated by the data-centric view in PGAS models. This design leverages a combination of highly tuned MPI two-sided semantics and an automatic, user-transparent split of MPI communicators to provide asynchronous progress. We implement the asynchronous progress ranks approach and other approaches within the Communication Runtime for Exascale which is a communication subsystem for Global Arrays. Our performance evaluation spans pure communication benchmarks, graph community detection and sparse matrix-vector multiplication kernels, and a computational chemistry application. The utility of our proposed PR-based approach is demonstrated by a 2.17x speed-up on 1008 processors over the other MPI-based designs.

Computational modeling↗

DOE OSTI2021

Cross Validation of the Network of Ground-Based Radar with GPM during the Remote Sensing of Electrification, Lightning, And Mesoscale/Microscale Processes with Adaptive Ground Observations (RELAMPAGO) Field Campaign

The cross-validation of radars in a network is important in making consistent retrievals across the domain and assuring the product quality. During the RELAMPAGO field campaign, two C-band radars, namely the Colorado State University C-band Hydrological Instrument for Volumetric Observations ( CSU-CHIVO) and the C-band Scanning ARM Precipitation Radar (CSAPR-2), were deployed near the Sierras de Cordoba in Argentina, a region known for having some of the most intense severe weather in the world. In addition to these two radars, the operational radar of the Cordoba City, the Radar Meteorologico Argentino 1 (RMA-1), adds another instrument to the RELAMPAGO network. This study presents an intercomparison between the RELAMPAGO C-band radars using the GPM spaceborne radar as a common reference ..A method to bring ground-based radars into better agreement is also proposed. Moreover, the attenuation correction for the C-band radar is studied in the context of intercomparing two radars. The attenuation coefficients are computed for the RELAMPAGO domain using the local disdrometers deployed. during the campaign. After the attenuation correction, CSU-CHIVO, CSAPR-2, and RMA-1 compare well with GPM-DPR with a high correlation and bias less than 1 dB.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2011

Corrective Action Decision Document/Closure Report for Corrective Action Unit 561: Waste Disposal Areas, Nevada National Security Site, Nevada, Revision 0 with ROTC 1

CAU 561 comprises 10 CASs: (1) 01-19-01, Waste Dump; (2) 02-08-02, Waste Dump and Burn Area; (3) 03-19-02, Debris Pile; (4) 05-62-01, Radioactive Gravel Pile; (5) 12-23-09, Radioactive Waste Dump; (6) 22-19-06, Buried Waste Disposal Site; (7) 23-21-04, Waste Disposal Trenches ; (8) 25-08-02, Waste Dump; (9) 25-23-21, Radioactive Waste Dump; and (10) 25-25-19, Hydrocarbon Stains and Trench. The purpose of this CADD/CR is to provide justification and documentation supporting the recommendation for closure of CAU 561 with no further corrective action. The purpose of the CAI was to fulfill the following data needs as defined during the DQO process: (1) Determine whether COCs are present; (2) If COCs are present, determine their nature and extent; and (3) Provide sufficient information and data to complete appropriate corrective actions. The following contaminants were determined to be present at concentrations exceeding their corresponding FALs: (1) No contamination exceeding FALs was identified at CASs 01-19-01, 03-19-02, 05-62-01, 12-23-09, and 22-19-06. (2) The surface and subsurface soil within the burn area at CAS 02-08-02 contains arsenic and lead above the FALs of 23 milligrams per kilogram (mg/kg) and 800 mg/kg, respectively. The surface and subsurface soil within the burn area also contains melted lead slag (potential source material [PSM]). The soil within the waste piles contains polyaromatic hydrocarbons (PAHs) above the FALs. The contamination within the burn area is spread throughout the area, as it was not feasible to remove all the PSM (melted lead), while at the waste piles, the contamination is confined to the piles. (3) The surface and subsurface soils within Trenches 3 and 5 at CAS 23-21-04 contain arsenic and polychlorinated biphenyls (PCBs) above the FALs of 23 mg/kg and 0.74 mg/kg, respectively. The soil was removed from both trenches, and the soil that remains at this CAS does not contain contamination exceeding the FALs. Lead bricks and counterweights were also removed, and the soil below these items does not contain contamination that exceeds the FAL for lead. (4) The concrete-like material at CAS 25-08-02 contains arsenic above the FAL of 23 mg/kg. This concrete-like material was removed, and the soil that remains at this CAS does not contain contamination exceeding the FALs. Lead-acid batteries were also removed, and the soil below the batteries does not contain contamination that exceeds the FAL for lead. (5) The surface soils within the main waste dump at the posted southern radioactive material area (RMA) at CAS 25-23-21 contain cesium (Cs)-137 and PCBs above the FALs of 72.9 picocuries per gram (pCi/g) and 0.74 mg/kg, respectively. The soil was removed from the RMA, and the soil that remains at this CAS does not contain contamination exceeding the FALs. (6) The surface and subsurface soils at CAS 25-25-19 do not contain contamination exceeding the FALs. In addition, lead bricks were removed, and the soil below these items does not contain contamination that exceeds the FAL for lead. The following best management practices were implemented: (1) Housekeeping debris at CASs 02-08-02, 23-21-04, 25-08-02, 25-23-21, and 25-25-19 was removed and disposed of; (2) The open trenches at CAS 23-21-04 were backfilled; (3) The waste piles at CAS 25-08-02 were removed and the area leveled to ground surface; and (4) The remaining waste piles at the main waste dump at CAS 25-23-21 were leveled to ground surface. Therefore, NNSA/NSO provides the following recommendations: (1) No further action for CASs 01-19-01, 03-19-02, 05-62-01, 12-23-09, and 22-19-06; (2) Closure in place with an FFACO use restriction (UR) at CAS 02-08-02 for the remaining PAH-, arsenic-, and lead-contaminated soil, and the melted lead PSM. The UR form and map have been filed in the NNSA/NSO Facility Information Management System, the FFACO database, and the NNSA/NSO CAU/CAS files; (3) No further corrective action at CAS 23-21-04, as the lead bricks and counterweights (PSM) have been removed, and the COCs of arsenic and PCBs in soil have been removed; (4) No further corrective action at CAS 25-08-02, as the COC of arsenic in soil has been removed, and the lead-acid batteries have been removed; (5) No further corrective action at CAS 25-23-21, as the COCs of Cs-137 and PCBs in soil have been removed, and the cast-iron pipes have been removed and disposed of; (6) No further corrective action at CAS 25-25-19, as the lead bricks (PSM) been removed; (7) A Notice of Completion to the NNSA/NSO is requested from the Nevada Division of Environmental Protection for closure of CAU 561; and (8) Corrective Action Unit 561 should be moved from Appendix III to Appendix IV of the FFACO.

12 MANAGEMENT OF RADIOACTIVE AND NON-RADIOACTIVE W↗

DOE OSTI2013

Corrective Action Investigation Plan for Corrective Action Unit 567: Miscellaneous Soil Sites, Nevada National Security Site, Nevada, with ROTC 1 Revision 0

Corrective Action Unit (CAU) 567 is located in Areas 1, 3, 5, 20, and 25 of the Nevada National Security Site, which is approximately 65 miles northwest of Las Vegas, Nevada. CAU 567 is a grouping of sites where there has been a suspected release of contamination associated with nuclear testing. This document describes the planned investigation of CAU 567, which comprises the following corrective action sites (CASs): • 01-23-03, Atmospheric Test Site T-1 • 03-23-25, Seaweed E Contamination Area • 05-23-07, A5b RMA • 20-23-08, Colby Mud Spill • 25-23-23, J-11 Soil RMA These sites are being investigated because existing information on the nature and extent of potential contamination is insufficient to evaluate and recommend corrective action alternatives (CAAs). Additional information will be obtained by conducting a corrective action investigation before evaluating CAAs and selecting the appropriate corrective action for each CAS. The results of the field investigation will support a defensible evaluation of viable CAAs that will be presented in the investigation report. The sites will be investigated based on the data quality objectives (DQOs) developed on May 6, 2013, by representatives of the Nevada Division of Environmental Protection and the U.S. Department of Energy (DOE), National Nuclear Security Administration Nevada Field Office. The DQO process was used to identify and define the type, amount, and quality of data needed to develop and evaluate appropriate corrective actions for CAU 567. The site investigation process will also be conducted in accordance with the Soils Activity Quality Assurance Plan, which establishes requirements, technical planning, and general quality practices to be applied to this activity. The potential contamination sources associated with CAU 567 releases are nuclear test operations and other NNSS operations. The DQO process resulted in an assumption that total effective dose (TED) within a default contamination boundary at Atmospheric Test Site T-1 exceeds the final action level (FAL) and requires corrective action. The presence and nature of contamination outside the default contamination boundary at Atmospheric Test Site T-1 and all other CAU 567 CASs will be evaluated based on information collected from a field investigation. Radiological contamination will be evaluated based on a comparison of the TED at sample locations to the dose-based FAL. The TED will be calculated as the total of separate estimates of internal and external dose. Results from the analysis of soil samples will be used to calculate internal radiological dose. Thermoluminescent dosimeters placed at the center of each sample location will be used to measure external radiological dose. Appendix A provides a detailed discussion of the DQO methodology and the DQOs specific to each CAS.

54 ENVIRONMENTAL SCIENCES↗

DOE OSTI2015

Materials compatibility and aging for flux and cleaner combinations

A materials study of high reliability electronics cleaning is presented here. In Phase 1, mixed type substrates underwent a condensed contaminants application to view a worst- case scenario for unremoved flux with cleaning agent residue for parts in a silicone oil filled environment. In Phase 2, fluxes applied to copper coupons and to printed wiring boards underwent gentle cleaning then accelerated aging in air at 65% humidity and 30 O C. Both sets were aged for 4 weeks. Contaminants were no-clean (ORL0), water soluble (ORH1 liquid and ORH0 paste), and rosin (RMA; ROL0) fluxes. Defluxing agents were water, solvents, and engineered aqueous defluxers. In the first phase, coupons had flux applied and heated, then were placed in vials of oil with a small amount of cleaning agent and additional coupons. In the second phase, pairs of copper coupons and PWB were hand soldered by application of each flux, using tin-lead solder in a strip across the coupon or a set of test components on the PWB. One of each pair was cleaned in each cleaning agent, the first with a typical clean, and the second with a brief clean. Ionic contamination residue was measured before accelerated aging. After aging, substrates were removed and a visual record of coupon damage made, from which a subjective rank was applied for comparison between the various flux and defluxer combinations; more corrosion equated to higher rank. The ORH1 water soluble flux resulted in the highest ranking in both phases, the RMA flux the least. For the first phase, in which flux and defluxer remained on coupons, the aqueous defluxers led to worse corrosion. The vapor phase cleaning agents resulted in the highest ranking in the second phase, in which there was no physical cleaning. Further study of cleaning and rinsing parameters will be required.

36 MATERIALS SCIENCE↗

DOE OSTI2016

Challenges of Monte Carlo Transport

These are slides from a presentation for Parallel Summer School at Los Alamos National Laboratory. Solving discretized partial differential equations (PDEs) of interest can require a large number of computations. We can identify concurrency to allow parallel solution of discrete PDEs. Simulated particles histories can be used to solve the Boltzmann transport equation. Particle histories are independent in neutral particle transport, making them amenable to parallel computation. Physical parameters and method type determine the data dependencies of particle histories. Data requirements shape parallel algorithms for Monte Carlo. Then, Parallel Computational Physics and Parallel Monte Carlo are discussed and, finally, the results are given. The mesh passing method greatly simplifies the IMC implementation and allows simple load-balancing. Using MPI windows and passive, one-sided RMA further simplifies the implementation by removing target synchronization. The author is very interested in implementations of PGAS that may allow further optimization for one-sided, read-only memory access (e.g. Open SHMEM). The MPICH_RMA_OVER_DMAPP option and library is required to make one-sided messaging scale on Trinitite - Moonlight scales poorly. Interconnect specific libraries or functions are likely necessary to ensure performance. BRANSON has been used to directly compare the current standard method to a proposed method on idealized problems. The mesh passing algorithm performs well on problems that are designed to show the scalability of the particle passing method. BRANSON can now run load-imbalanced, dynamic problems. Potential avenues of improvement in the mesh passing algorithm will be implemented and explored. A suite of test problems that stress DD methods will elucidate a possible path forward for production codes.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗

DOE OSTI2018

GASNet-EX Performance Improvements Due to Specialization for the Cray Aries Network

This document is a deliverable for milestone STPM17-6 of the Exascale Computing Project, delivered by WBS 2.3.1.14. It reports on the improvements in performance observed on Cray XC-series systems due to enhancements made to the GASNet-EX software. These enhancements, known as “specializations”, primarily consist of replacing network-independent implementations of several recently added features with implementations tailored to the Cray Aries network. Performance gains from specialization include (1) Negotiated-Payload Active Messages improve bandwidth of a ping-pong test by up to 14%, (2) Immediate Operations reduce running time of a synthetic benchmark by up to 93%, (3) non-bulk RMA Put bandwidth is increased by up to 32%, (4) Remote Atomic performance is 70% faster than the reference on a point-to-point test and allows a hot-spot test to scale robustly, and (5) non-contiguous RMA interfaces see up to 8.6x speedups for an intra-node benchmark and 26% for inter-node. These improvements are available in the GASNet-EX 2018.3.0 release.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2023

UPC++ v1.0 Programmer’s Guide (Rev. 2023.9.0)

UPC++ is a C++ library that supports Partitioned Global Address Space (PGAS) programming. It is designed for writing efficient, scalable parallel programs on distributed-memory parallel computers. The key communication facilities in UPC++ are one-sided Remote Memory Access (RMA) and Remote Procedure Call (RPC). The UPC++ control model is single program, multiple-data (SPMD), with each separate constituent process having access to local memory as it would in C++. The PGAS memory model additionally provides one-sided RMA communication to a global address space, which is allocated in shared segments that are distributed over the processes. UPC++ also features Remote Procedure Call (RPC) communication, making it easy to move computation to operate on data that resides on remote processes. UPC++ was designed to support exascale high-performance computing, and the library interfaces and implementation are focused on maximizing scalability. In UPC++, all communication operations are syntactically explicit, which encourages programmers to consider the costs associated with communication and data movement. Moreover, all communication operations are asynchronous by default, encouraging programmers to seek opportunities for overlapping communication latencies with other useful work. UPC++ provides expressive and composable abstractions designed for efficiently managing aggressive use of asynchrony in programs. Together, these design principles are intended to enable programmers to write applications using UPC++ that perform well even on hundreds of thousands of cores.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2024

Implementing One Sided Partitioned Communication in Open MPI

This report introduces partitioned communication, a new MPI 4.0 interface that enables early bird communication by overlapping communication and computation. By partitioning messages into smaller sub-messages, MPI can start partial data transfers early. Performance studies show that the RMA implementation outperforms the Persistent implementation, despite some constraints. This report details a new opt-in RMA implementation, offering a high-performance option for partitioned communication that imposes some additional limitations.

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2022

UPC++ v1.0 Programmer’s Guide (Revision 2022.3.0)

97 MATHEMATICS AND COMPUTING↗

DOE OSTI2021

Optimization of Asynchronous Communication Operations through Eager Notifications

UPC++ is a C++ library implementing the Asynchronous Partitioned Global Address Space (APGAS) model. We propose an enhancement to the completion mechanisms of UPC++ used to synchronize communication operations that is designed to reduce overhead for on-node operations. Our enhancement permits eager delivery of completion notification in cases where the data transfer semantics of an operation happen to complete synchronously, for example due to the use of shared-memory bypass. This semantic relaxation allows removing significant overhead from the critical path of the implementation in such cases. We evaluate our results on three different representative systems using a combination of microbenchmarks and five variations of the the HPCChallenge RandomAccess benchmark implemented in UPC++ and run on a single node to accentuate the impact of locality. We find that in RMA versions of the benchmark written in a straightforward manner (without manually optimizing for locality), the new eager notification mode can provide up to a 25% speedup when synchronizing with promises and up to a 13.5x speedup when synchronizing with conjoined futures. We also evaluate our results using a graph matching application written with UPC++ RMA communication, where we measure overall speedups of as much as 11% in single-node runs of the unmodified application code, due to our transparent enhancements.

Kamil, Amir↗