A framework for characterizing overlap of communication and computation in parallel applications

Effective overlap of computation and communication is a well understood technique for latency hiding and can yield significant performance gains for applications on high-end computers. In this paper, we propose an instrumentation framework for message-passing systems to characterize the degree of overlap of communication with computation in the execution of parallel applications. The inability to obtain precise time-stamps for pertinent communication events is a significant problem, and is addressed by generation of minimum and maximum bounds on achieved overlap. The overlap measures can aid application developers and system designers in investigating scalability issues. The approach has been used to instrument two MPI implementations as well as the ARMCI system. The implementation resides entirely within the communication library and thus integrates well with existing approaches that operate outside the library. The utility of the framework is demonstrated by analyzing communication-computation overlap for micro-benchmarks and the NAS benchmarks, and the insights obtained are used to modify the NAS SP benchmark, resulting in improved overlap.

Environmental benchmarks for buildings: a critical literature review

Purpose

To reduce the environmental impact of the building sector, environmental targets considering the full life cycle of buildings can be supportive. In recent years, various benchmarks based on Life Cycle Assessment (LCA) have been developed as part of regulations, labelling systems, sustainability rating tools and research studies. The objective of this paper is to critically analyse 23 existing benchmarking systems focusing on the benchmark methodology but also on the benchmark applications and communication.

Methods

The critical literature review consists of two parts. In a first part, the choices related to the assessment method, functional equivalent, definition of benchmark values, benchmark scope, benchmark applications and benchmark communication are compared. In the second part, benchmark values are compiled from literature and statistically analysed.

Results and discussion

The comparative analysis allows to identify the main approaches and methods used in benchmarking systems. For each evaluation aspect, the strengths and weaknesses of the various approaches are highlighted. The statistical analysis provides insight in the spread of benchmark values. Important variations are found between the literature sources which can be explained by differences in benchmark approach, scope, system boundaries and applications.

Conclusions

Based on the comparative analysis, recommendations are formulated for the development of LCA benchmarks for the building sector. The results of the statistical analysis furthermore provide reference values which can be used for the validation of future benchmarks. For global warming, the statistical values for the full life cycle impacts (i.e. embodied and operational impacts) range from about 15 up to 35 kg CO₂ eq/m².a.

     

OXBench: A benchmark for evaluation of protein multiple sequence alignment accuracy

Background  

The alignment of two or more protein sequences provides a powerful guide in the prediction of the protein structure and in identifying key functional residues, however, the utility of any prediction is completely dependent on the accuracy of the alignment. In this paper we describe a suite of reference alignments derived from the comparison of protein three-dimensional structures together with evaluation measures and software that allow automatically generated alignments to be benchmarked. We test the OXBench benchmark suite on alignments generated by the AMPS multiple alignment method, then apply the suite to compare eight different multiple alignment algorithms. The benchmark shows the current state-of-the art for alignment accuracy and provides a baseline against which new alignment algorithms may be judged.     

Benchmarking of dynamic simulation predictions in two software platforms using an upper limb musculoskeletal model

Several opensource or commercially available software platforms are widely used to develop dynamic simulations of movement. While computational approaches are conceptually similar across platforms, technical differences in implementation may influence output. We present a new upper limb dynamic model as a tool to evaluate potential differences in predictive behavior between platforms. We evaluated to what extent differences in technical implementations in popular simulation software environments result in differences in kinematic predictions for single and multijoint movements using EMG- and optimization-based approaches for deriving control signals. We illustrate the benchmarking comparison using SIMM–Dynamics Pipeline–SD/Fast and OpenSim platforms. The most substantial divergence results from differences in muscle model and actuator paths. This model is a valuable resource and is available for download by other researchers. The model, data, and simulation results presented here can be used by future researchers to benchmark other software platforms and software upgrades for these two platforms.     

Automated methods of predicting the function of biological sequences using GO and BLAST

Background  

With the exponential increase in genomic sequence data there is a need to develop automated approaches to deducing the biological functions of novel sequences with high accuracy. Our aim is to demonstrate how accuracy benchmarking can be used in a decision-making process evaluating competing designs of biological function predictors. We utilise the Gene Ontology, GO, a directed acyclic graph of functional terms, to annotate sequences with functional information describing their biological context. Initially we examine the effect on accuracy scores of increasing the allowed distance between predicted and a test set of curator assigned terms. Next we evaluate several annotator methods using accuracy benchmarking. Given an unannotated sequence we use the Basic Local Alignment Search Tool, BLAST, to find similar sequences that have already been assigned GO terms by curators. A number of methods were developed that utilise terms associated with the best five matching sequences. These methods were compared against a benchmark method of simply using terms associated with the best BLAST-matched sequence (best BLAST approach).     

Improved MPI collectives for MPI processes in shared address spaces

As the number of cores per node keeps increasing, it becomes increasingly important for MPI to leverage shared memory for intranode communication. This paper investigates the design and optimization of MPI collectives for clusters of NUMA nodes. We develop performance models for collective communication using shared memory and we demonstrate several algorithms for various collectives. Experiments are conducted on both Xeon X5650 and Opteron 6100 InfiniBand clusters. The measurements agree with the model and indicate that different algorithms dominate for short vectors and long vectors. We compare our shared-memory allreduce with several MPI implementations—Open MPI, MPICH2, and MVAPICH2—that utilize system shared memory to facilitate interprocess communication. On a 16-node Xeon cluster and 8-node Opteron cluster, our implementation achieves on geometric average 2.3X and 2.1X speedup over the best MPI implementation, respectively. Our techniques enable an efficient implementation of collective operations on future multi- and manycore systems.     

Protein models: The Grand Challenge of protein docking

Characterization of life processes at the molecular level requires structural details of protein–protein interactions (PPIs). The number of experimentally determined protein structures accounts only for a fraction of known proteins. This gap has to be bridged by modeling, typically using experimentally determined structures as templates to model related proteins. The fraction of experimentally determined PPI structures is even smaller than that for the individual proteins, due to a larger number of interactions than the number of individual proteins, and a greater difficulty of crystallizing protein–protein complexes. The approaches to structural modeling of PPI (docking) often have to rely on modeled structures of the interactors, especially in the case of large PPI networks. Structures of modeled proteins are typically less accurate than the ones determined by X‐ray crystallography or nuclear magnetic resonance. Thus the utility of approaches to dock these structures should be assessed by thorough benchmarking, specifically designed for protein models. To be credible, such benchmarking has to be based on carefully curated sets of structures with levels of distortion typical for modeled proteins. This article presents such a suite of models built for the benchmark set of the X‐ray structures from the Dockground resource ( http://dockground.bioinformatics.ku.edu ) by a combination of homology modeling and Nudged Elastic Band method. For each monomer, six models were generated with predefined C^α root mean square deviation from the native structure (1, 2, …, 6 Å). The sets and the accompanying data provide a comprehensive resource for the development of docking methodology for modeled proteins. Proteins 2014; 82:278–287. © 2013 Wiley Periodicals, Inc.     

Frustrations in benchmarking IRBs: reflections after analyzing the United States' IRB registration database

Like any organization committed to quality, institutional review boards (IRBs) desire to benchmark themselves with their colleagues. In order to do so, there is a need for high-quality, well-validated, and well-populated datasets designed for this purpose. Unlike other areas of healthcare, there are no such aggregated databases of IRB information. Analyzing the largest repository of IRB information--the United States' IRB registration database consisting of approximately 4,500 IRBs--yields some useful results, but given that this database was designed as a registry and not for benchmarking, its use is limited to benchmarking the most basic demographics. If members of the IRB community want to benchmark themselves relative to their colleagues, they will have to look at how other healthcare operations have successfully done it (through private and/or government efforts) and invest the time and resources required to obtain meaningful data.     

Coronary computed tomography angiography and [15O]H2O positron emission tomography perfusion imaging for the assessment of coronary artery disease

Determining the anatomic severity and extent of coronary artery disease (CAD) by means of coronary computed tomography angiography (CCTA) and its effect on perfusion using myocardial perfusion imaging (MPI) form the pillars of the non-invasive imaging assessment of CAD. This review will 1) focus on CCTA and [¹⁵O]H₂O positron emission tomography MPI as stand-alone imaging modalities and their combined use for detecting CAD, 2) highlight some of the lessons learned from the PACIFIC trial (Comparison of Coronary CT Angiography, SPECT, PET, and Hybrid Imaging for Diagnosis of Ischemic Heart Disease Determined by Fractional Flow Reserve (FFR) (NCT01521468)), and 3) discuss the use of [¹⁵O]H₂O PET MPI in the clinical work-up of patients with a chronic coronary total occlusion (CTO).

     

The design and implementation of MPI collective operations for clusters in long-and-fast networks

Several MPI systems for Grid environment, in which clusters are connected by wide-area networks, have been proposed. However, the algorithms of collective communication in such MPI systems assume relatively low bandwidth wide-area networks, and they are not designed for the fast wide-area networks that are becoming available. On the other hand, for cluster MPI systems, a bcast algorithm by van de Geijn, et al. and an allreduce algorithm by Rabenseifner have been proposed, which are efficient in a high bi-section bandwidth environment. We modify those algorithms so as to effectively utilize fast wide-area inter-cluster networks and to control the number of nodes which can transfer data simultaneously through wide-area networks to avoid congestion. We confirmed the effectiveness of the modified algorithms by experiments using a 10 Gbps emulated WAN environment. The environment consists of two clusters, where each cluster consists of nodes with 1 Gbps Ethernet links and a switch with a 10 Gbps upper link. The two clusters are connected through a 10 Gbps WAN emulator which can insert latency. In a 10 millisecond latency environment, when the message size is 32 MB, the proposed bcast and allreduce are 1.6 and 3.2 times faster, respectively, than the algorithms used in existing MPI systems for Grid environment.

Rapid Simulation of Unprocessed DEER Decay Data for Protein Fold Prediction

Despite advances in sampling and scoring strategies, Monte Carlo modeling methods still struggle to accurately predict de novo the structures of large proteins, membrane proteins, or proteins of complex topologies. Previous approaches have addressed these shortcomings by leveraging sparse distance data gathered using site-directed spin labeling and electron paramagnetic resonance spectroscopy to improve protein structure prediction and refinement outcomes. However, existing computational implementations entail compromises between coarse-grained models of the spin label that lower the resolution and explicit models that lead to resource-intense simulations. These methods are further limited by their reliance on distance distributions, which are calculated from a primary refocused echo decay signal and contain uncertainties that may require manual refinement. Here, we addressed these challenges by developing RosettaDEER, a scoring method within the Rosetta software suite capable of simulating double electron-electron resonance spectroscopy decay traces and distance distributions between spin labels fast enough to fold proteins de novo. We demonstrate that the accuracy of resulting distance distributions match or exceed those generated by more computationally intensive methods. Moreover, decay traces generated from these distributions recapitulate intermolecular background coupling parameters even when the time window of data collection is truncated. As a result, RosettaDEER can discriminate between poorly folded and native-like models by using decay traces that cannot be accurately converted into distance distributions using regularized fitting approaches. Finally, using two challenging test cases, we demonstrate that RosettaDEER leverages these experimental data for protein fold prediction more effectively than previous methods. These benchmarking results confirm that RosettaDEER can effectively leverage sparse experimental data for a wide array of modeling applications built into the Rosetta software suite.     

Evaluating interspecific niche overlaps in environmental and geographic spaces to assess the value of umbrella species

The concept of umbrella species assumes that concentrating resources on the protection of a single species contributes to the conservation of a suite of species and ecological processes belonging to the same ecosystem. The environmental requirements and geographical distribution of the umbrella species should thus overlap those of the group of targeted species. In western France, the conservation of several large grassland floodplains relies on agri‐environmental schemes targeting one single bird species, the corncrake Crex crex. It is considered as an umbrella species but no real assessment of its effectiveness has been carried out so far. We used a two‐step methodology to assess the potential of the corncrake to act as an umbrella species by estimating niche overlap in the environmental and geographic space between the main ground‐nesting species of the bird community in these grasslands, including the corncrake and four passerines. The five species showed substantial differences in their ecological niches so that their distributions did not perfectly overlap. Overlaps in predicted distributions between pairs of species depended on the threshold used to convert suitability to binary maps. Moreover, the number of species that could be protected by a candidate umbrella species was affected by the overlap criterion of success. Although the corncrake may be used as an umbrella species, it would be outperformed by several passerine species. Our study highlights the potential of using niche overlap to select umbrella species. It also reveals the importance of analysing the sensitivity of outputs when changing thresholds and overlap criteria.     

Diets of sympatric native and introduced carnivores in the Barrington Tops,eastern Australia

Invasive predators have severe impacts on global biodiversity, and their effects in Australia have been more extreme than on any other continent. The spotted‐tailed quoll (Dasyurus maculatus), an endangered marsupial carnivore, coexists with three eutherian carnivores, the red fox (Vulpes vulpes), feral cat (Felis catus) and wild dog (Canis lupus ssp.) with which it did not coevolve. No previous study has investigated dietary overlap between quolls and the suite of three eutherian carnivores. By analysing scats, we aimed to quantify dietary overlap within this carnivore assemblage in eastern Australia, and to detect any differences that may facilitate coexistence. We also sought evidence of intraguild predation. Dietary overlap between predators was extensive, with the greatest similarity occurring between foxes and cats. However, some differences were apparent. For example, cats mainly consumed smaller prey, and wild dogs larger prey. Quolls showed greater dietary overlap with foxes and cats than with dogs. Intraguild predation was evident, with fox remains occurring in 3% of wild dog scats. Our results suggest wild dogs competitively dominate invasive foxes, which in turn are likely to compete with the endangered quoll.     

Observational constraints reduce model spread but not uncertainty in global wetland methane emission estimates

The recent rise in atmospheric methane (CH₄) concentrations accelerates climate change and offsets mitigation efforts. Although wetlands are the largest natural CH₄ source, estimates of global wetland CH₄ emissions vary widely among approaches taken by bottom-up (BU) process-based biogeochemical models and top-down (TD) atmospheric inversion methods. Here, we integrate in situ measurements, multi-model ensembles, and a machine learning upscaling product into the International Land Model Benchmarking system to examine the relationship between wetland CH₄ emission estimates and model performance. We find that using better-performing models identified by observational constraints reduces the spread of wetland CH₄ emission estimates by 62% and 39% for BU- and TD-based approaches, respectively. However, global BU and TD CH₄ emission estimate discrepancies increased by about 15% (from 31 to 36 TgCH₄ year⁻¹) when the top 20% models were used, although we consider this result moderately uncertain given the unevenly distributed global observations. Our analyses demonstrate that model performance ranking is subject to benchmark selection due to large inter-site variability, highlighting the importance of expanding coverage of benchmark sites to diverse environmental conditions. We encourage future development of wetland CH₄ models to move beyond static benchmarking and focus on evaluating site-specific and ecosystem-specific variabilities inferred from observations.     

Benchmarking academic plastic surgery services in the United States

Rising health care costs and increasingly demanding patients are only some of the challenges faced by academic plastic surgery services in their pursuit of excellence in education, research, and patient care. Benchmarking, when correctly applied, is a powerful tool that can help services learn from each other's experiences. This study aimed at creating the first benchmarking report summarizing performance indicators and management practices of some of the most complete academic plastic surgery units in the United States. Results provide an opportunity for plastic surgery leaders to benchmark against their own units, identify eventual gaps, and improve their performance as needed.     

Ischémie myocardique du diabétique : apports de la scintigraphie myocardique de perfusion

Diabetes is an important risk factor of myocardial ischemia. Myocardial perfusion imaging (MPI) is used for diagnostic, prognostic assessment and for post-therapeutic follow-up of coronary artery disease (CAD). Its usefulness has been documented extensively in the general population. However, in diabetic population, little studies have been published. CAD is more severe and more frequent among diabetic patients. The goal of this work was to assess usefulness of MPI among diabetic patients. This work includes 44 diabetic patients (22 women, 22 men), investigated by stress MPI in nuclear medicine department of Ibn Rochd UH of Casablanca. The studied parameters concerned: age, antecedents of CAD, risk factors of CAD associated to diabetes, duration of the diabetes, diabetes complications, diabetes treatment, indication of MPI, presence or not of anomaly on the ECG performed at rest, existence or not of typical or atypical clinical signs of CAD, investigations and therapy prescribed as well as the occurrence or not of cardiac event during monitoring after MPI. Results of MPI have been compared to clinical, therapeutic and monitoring data of patients. Mean age of patients was 55 years (39 to 75 years), mean diabetes duration was 8.6 years (1 to 30 years), at least one diabetes complication has been noted in 18 patients. The most frequent complication was diabetic retinopathy. MPI has been achieved for diagnosis of ischemia in 37 patients and assessment of anti-ischemic treatment in 7 cases. Treadmill exercise has been achieved in 34 cases and a pharmacological stress in 10 others. During follow-up, which was between 1 and 36 months (mean: 14.9 months), 6 cardiac events occurred among the 44 patients. Patients with abnormal findings at stress MPI had two cardiovascular risk factors or more associated to diabetes (91.3% vs. 9.5% among patients having normal findings, P < 0.001) and had diabetic retinopathy more often (56.5% vs. 9.5% among patients with normal findings, P < 0.001). Cardiac events were more frequent among men (100% vs. 43.2% of patients who did not have a coronary event, P < 0.01). Patients with stress MPI showing ischemia in 3/17 segments or more have presented a cardiac event more often during the follow-up (4/12 Vs 2/32 among patients with normal MPI or defect in less than 3/17 segments, P < 0.01). In this series, coronary artery disease was found more frequently among patients having more than 2 cardiovascular risk factors. In addition, risk of cardiac event seems related to extent of uptake decrease.     

Patterns and partitioning of food resources by elasmobranchs in southern Brazil

Our study sought to analyze patterns and overlap in prey items consumed by elasmobranch species in southern Brazil through a complex network analysis. We hypothesized that species with the same habitat use would show greater trophic niche overlap and lower diet specialization than species with different habitat use. We conducted a literature survey of the species recorded in southern Brazil and analyzed publications with quantitative data on the diet of each species. We found records of 103 species in the region, of which 58 had their diets analyzed and had quantitative data to contribute to this analysis. The results suggest nested diet patterns in this community. Nestedness may arise for different reasons, and suggests that species might be balancing their high feeding niche overlap through compensatory mechanisms. There are no prior records of such a pattern in elasmobranch communities, which brings to light important information about their feeding niche. Data on prey consumption and temporal and spatial patterns of elasmobranch feeding may yield further insight into how these species are interacting in the environment and assist the investigation of the processes that shaped the current community’s trophic structure.

     

Performances diagnostiques des tomoscintigraphies myocardiques de stress,réalisées avec une caméra à semi-conducteurs et un protocole très faible dose

IntroductionThis retrospective study aimed to assess the performances in the detection of coronary artery disease patients of the stress myocardial perfusion imaging (MPI) performed with a semiconductor camera, a very low dose stress-rest protocol and short recording times.Material and methodsWe analyzed consecutive MPI exams, which has been routinely planed with the “D-SPECT” semiconductor camera, a 1-day stress-rest protocol, very low doses of Sestamibi-^99mTc (120 MBq at stress and 360 MBq at rest for a 75 kg body weight patient) and short acquisition-times targeting the recording of 500 myocardial kcounts (on average, 8 min for stress and 3 min for rest). The ability to detect coronary artery stenosis (> 50% diameter reduction) was assessed in a group of 118 patients, who had coronary angiography at ≤ 3 months from MPI, and normalcy rate was assessed in a group of 74 patients showing a low pretest probability of coronary artery disease (< 10%).ResultsSensitivity, positive predictive value and global accuracy for identifying the 81 patients with ≥ 1 coronary artery stenosis were 85%, 83% and 78%, respectively; and normalcy rate was 96%. Mean effective doses were: (1) 4.9 ± 1.4 mSv in the group with coronary angiography and where most exams (90%) comprised both stress and rest MPI; and (2) 1.9 ± 1.5 mSv in the low probability group and where most exams (70%) comprised only stress MPI.ConclusionWhen performed with a sensitive semiconductor camera and a very low dose stress-rest protocol, MPI provides high diagnostic performances, equivalent to those documented with conventional cameras in the same study setting, but with dramatic reduction of patients’ radiation.     

Altitudinal distribution of the entire invasive small mammal guild in the eastern dryland zone of New Zealand’s Southern Alps

New Zealand faces significant ecological problems caused by the introduction of a variety of invasive small mammal species. Many of these species originate from temperate to subarctic climates and occur across wide elevations in their native range, and so arrived predisposed to adapt to a variety of habitats and bioclimatic zones in their new environs, including the alpine zone. Almost all of New Zealand’s invasive small mammal species have been recorded in the country’s alpine zones, yet neither the altitudinal distribution nor the extent to which such species use high elevation areas has been clearly defined. We conducted extensive camera trap surveys in summer to autumn periods of 2019 and 2020 across an elevation range of 500–2250 m above sea level, and used detection rates and occupancy modelling to reveal the altitudinal distributions and habitat associations of all 10 invasive small mammals that occur in the dryland zone of the central South Island. We found altitudinal distributions varied greatly across species, and that while most exhibited decreasing detection rates and site occupancy probabilities with increasing elevation, some used the subalpine and alpine zones to a greater degree than adjacent lower elevations. There were clear habitat associations, as well as interspecific associations that helped to explain the altitudinal distribution of some species. Understanding how such factors influence the distribution of invasive small mammals has both broad implications for invasive species management, and direct applications in evaluating threats to native taxa, advancing management strategies, and benchmarking distributions in a changing climate.