With the increased prevalence of employing workflows for scientific computing and a push towards exascale computing, it has become paramount that we are able to analyze characteristics of scientific applications to better understand their impact on the underlying infrastructure and vice-versa. Such analysis can help drive the design, development, and optimization of these next generation systems and solutions. In this paper, we present the architecture, integrated with existing well-established and newly developed tools, to collect online performance statistics of workflow executions from various, heterogeneous sources and publish them in a distributed database (Elasticsearch). Using this architecture, we are able to correlate online workflow performance data, with data from the underlying infrastructure, and present them in a useful and intuitive way via an online dashboard. We have validated our approach by executing two classes of real-world workflows, both under normal and anomalous conditions. The first is an I/O-intensive genome analysis workflow; the second, a CPU- and memory-intensive material science workflow. Based on the data collected in Elasticsearch, we are able to demonstrate that we can correctly identify anomalies that we injected. The resulting end-to-end data collection of workflow performance data is an important resource of training data for automated machine learning analysis.

Crystallography is the powerhouse technique for molecular structure determination, with applications in fields ranging from energy storage to drug design. Accurate structure determination, however, relies partly on determining the precise locations and integrated intensities of Bragg peaks in the resulting data. Here, we describe a method for Bragg peak integration that is accomplished using neural networks. The network is based on a U-Net and identifies peaks in three-dimensional reciprocal space through segmentation, allowing prediction of the full 3D peak shape from noisy data that is commonly difficult to process. The procedure for generating appropriate training sets is detailed. Trained networks achieve Dice coefficients of 0.82 and mean IoUs of 0.69. Carrying out integration over entire datasets, it is demonstrated that integrating neural network-predicted peaks results in improved intensity statistics. Furthermore, using a second dataset, the possibility of transfer learning between datasets is shown. Given the ubiquity and growing complexity of crystallography, we anticipate integration by machine learning to play an increasingly important role across the physical sciences. These early results demonstrate the applicability of deep learning techniques for integrating crystallography data and suggest a possible role in the next generation of crystallography experiments.

Neutron crystallography offers enormous potential to complement structures from X-ray crystallography by clarifying the positions of low-Z elements, namely hydrogen. Macromolecular neutron crystallography, however, remains limited, in part owing to the challenge of integrating peak shapes from pulsed-source experiments. To advance existing software, this article demonstrates the use of machine learning to refine peak locations, predict peak shapes and yield more accurate integrated intensities when applied to whole data sets from a protein crystal. The artificial neural network, based on the U-Net architecture commonly used for image segmentation, is trained using about 100000 simulated training peaks derived from strong peaks. After 100 training epochs (a round of training over the whole data set broken into smaller batches), training converges and achieves a Dice coefficient of around 65%, in contrast to just 15% for negative control data sets. Integrating whole peak sets using the neural network yields improved intensity statistics compared with other integration methods, including k-nearest neighbours. These results demonstrate, for the first time, that neural networks can learn peak shapes and be used to integrate Bragg peaks. It is expected that integration using neural networks can be further developed to increase the quality of neutron, electron and X-ray crystallography data.

Neutron crystallography is a powerful technique for directly visualizing the locations of H atoms in biological macromolecules. This information has provided key new insights into enzyme mechanisms, ligand binding and hydration. However, despite the importance of this information, the application of neutron crystallography in biology has been limited by the relatively low flux of available neutron beams and the large incoherent neutron scattering from hydrogen, both of which contribute to weak diffraction data with relatively low signal-to-background ratios. A method has been developed to fit weak data based on three-dimensional profile fitting of Bragg peaks in reciprocal space by an Ikeda–Carpenter function with a bivariate Gaussian. When applied to data collected from three different proteins, three-dimensional profile fitting yields intensities with higher correlation coefficients (CC1/2) at high resolutions, decreased Rfree factors, extended resolutions and improved nuclear density maps. Importantly, additional features are revealed in nuclear density maps that may provide additional scientific information. These results suggest that three-dimensional profile fitting will help to extend the capabilities of neutron macromolecular crystallography.

This paper presents a method for performance profiles development of scientific workflow. It addresses issues related to: workflows execution in a parameter sweep manner, collecting performance information about each workflow task, and analysis of the collected data with statistical learning methods. The main goal of this work is to increase the understanding about the performance of studied workflows in a systematic and predictable way. The evaluation of the presented approach is based on a real scientific workflow developed by the Spallation Neutron Source - a DOE research facility at the Oak Ridge National Laboratory. The workflow executes an ensemble of molecular dynamics and neutron scattering intensity calculations to optimize a model parameter value.

Modeling helps explain the fundamental physics hidden behind experimental data. In the case of material modeling, running one simulation rarely results in output that reproduces the experimental data. Often one or more of the force field parameters are not precisely known and must be optimized for the output to match that of the experiment. Since the simulations require high performance computing (HPC) resources and there are usually many simulations to run, a workflow is very useful to prevent errors and assure that the simulations are identical except for the parameters that need to be varied. The use of HPC implies distributed workflows, but the optimization and steps to compare the simulation results and experimental data are done on a local workstation. We will present results from force field refinement of data collected at the Spallation Neutron Source using Kepler, Pegasus, and BEAM workflows and discuss what we have learned from using these workflows.

Large-scale simulations and data analysis are often required to explain neutron scattering experiments to establish a connection between the fundamental physics at the nanoscale and data probed by neutrons. However, to perform simulations at experimental conditions it is critical to use correct force-field (FF) parameters which are unfortunately not available for most complex experimental systems. In this work, we have developed a workflow optimization technique to provide optimized FF parameters by comparing molecular dynamics (MD) to neutron scattering data. We describe the workflow in detail by using an example system consisting of tRNA and hydrophilic nanodiamonds in a deuterated water (D2O) environment. Quasi-elastic neutron scattering (QENS) data show a faster motion of the tRNA in the presence of nanodiamond than without the ND. To compare the QENS and MD results quantitatively, a proper choice of FF parameters is necessary. We use an efficient workflow to optimize the FF parameters between the hydrophilic nanodiamond and water by comparing to the QENS data. Our results show that we can obtain accurate FF parameters by using this technique. The workflow can be generalized to other types of neutron data for FF optimization, such as vibrational spectroscopy and spin echo.

[This corrects the article DOI: 10.1107/S1600576716001369.].

Quasi-elastic neutron scattering (QENS) is one of the experimental techniques of choice for probing the dynamics at length and time scales that are also in the realm of full-atom molecular dynamics (MD) simulations. This overlap enables extension of current fitting methods that use time-independent equilibrium measurements to new methods fitting against dynamics data. We present an algorithm that fits simulation-derived incoherent dynamical structure factors against QENS data probing the diffusive dynamics of the system. We showcase the difficulties inherent to this type of fitting problem, namely, the disparity between simulation and experiment environment, as well as limitations in the simulation due to incomplete sampling of phase space. We discuss a methodology to overcome these difficulties and apply it to a set of full-atom MD simulations for the purpose of refining the force-field parameter governing the activation energy of methyl rotation in the octa-methyl polyhedral oligomeric silsesquioxane molecule. Our optimal simulated activation energy agrees with the experimentally derived value up to a 5% difference, well within experimental error. We believe the method will find applicability to other types of diffusive motions and other representation of the systems such as coarse-grain models where empirical fitting is essential. Also, the refinement method can be extended to the coherent dynamic structure factor with no additional effort.

Modern science is often conducted on large scale, distributed, heterogeneous and high-performance computing infrastructures. Increasingly, the scale and complexity of both the applications and the underlying execution platforms have been growing. Scientific workflows have emerged as a flexible representation to declaratively express complex applications with data andcontrol dependences. However, it is extremely challengingfor scientists to execute their science workflows in a reliable and scalable way due to a lack of understanding of expected and realistic behavior of complex scientific workflows on large scale and distributed HPC systems. This is exacerbated by failures and anomalies in largescale systems and applications, which makes detecting, analyzing and acting on anomaly events challenging. In this work, we present a prototype of an end-to-end system for modeling and diagnosing the runtime performance of complex scientific workflows. We interfaced the Pegasus workflow management system, Aspen performance modeling, monitoring and anomaly detection into an integrated framework that not only improves the understanding of complex scientific applications on large scale complex infrastructure, but also detects anomalies and supports adaptivity. We present a black box modeling tool, a comprehensive online monitoring system, and anomaly detection algorithms that employ the models and monitoring data to detect anomaly events. We present an evaluation of the system with a Spallation Neutron Source workflow as a driving use case.

Evidence is mounting that potentially exploitable properties of technologically and chemically interesting crystalline materials are often attributable to local structure effects, which can be observed as modulated diffuse scattering (mDS) next to Bragg diffraction (BD). BD forms a regular sparse grid of intense discrete points in reciprocal space. Traditionally, the intensity of each Bragg peak is extracted by integration of each individual reflection first, followed by application of the required corrections. In contrast, mDS is weak and covers expansive volumes of reciprocal space close to, or between, Bragg reflections. For a representative measurement of the diffuse scattering, multiple sample orientations are generally required, where many points in reciprocal space are measured multiple times and the resulting data are combined. The common post-integration data reduction method is not optimal with regard to counting statistics. A general and inclusive data processing method is needed. In this contribution, a comprehensive data analysis approach is introduced to correct and merge the full volume of scattering data in a single step, while correctly accounting for the statistical weight of the individual measurements. Development of this new approach required the exploration of a data treatment and correction protocol that includes the entire collected reciprocal space volume, using neutron time-of-flight or wavelength-resolved data collected at TOPAZ at the Spallation Neutron Source at Oak Ridge National Laboratory.

Errors in the paper by Michels-Clark, Lynch, Hoffmann, Hauser, Weber, Harrison & Bürgi [J. Appl. Cryst. (2013), 46, 1616–1625] are corrected.

Computational science is well established as the third pillar of scientific discovery and is on par with experimentation and theory. However, as we move closer toward the ability to execute exascale calculations and process the ensuing extreme-scale amounts of data produced by both experiments and computations alike, the complexity of managing the compute and data analysis tasks has grown beyond the capabilities of domain scientists. Thus, workflow management systems are absolutely necessary to ensure current and future scientific discoveries. A key research question for these workflow management systems concerns the performance optimization of complex calculation and data analysis tasks. The central contribution of this article is a description of the PANORAMA approach for modeling and diagnosing the run-time performance of complex scientific workflows. This approach integrates extreme-scale systems testbed experimentation, structured analytical modeling, and parallel systems simulation into a comprehensive workflow framework called Pegasus for understanding and improving the overall performance of complex scientific workflows.

The Mantid framework is a software solution developed for the analysis and visualization of neutron scattering and muon spin measurements. The framework is jointly developed by software engineers and scientists at the ISIS Neutron and Muon Facility and the Oak Ridge National Laboratory. The objectives, functionality and novel design aspects of Mantid are described.

Critical infrastructures have a number of the characteristic properties of complex systems. Among these are infrequent large failures through cascading events. These events, though infrequent, often obey a power law distribution in their probability versus size which suggests that conventional risk analysis does not apply to these systems. Real infrastructure systems typically have an additional layer of complexity, namely the heterogeneous coupling to other infrastructure systems that can allow a failure in one system to propagate to the other system. Here, we model the infrastructure systems through a network with complex system dynamics. We use both mean field theory to get analytic results and a numerical complex systems model, Demon, for computational results. An isolated system has bifurcated fixed points and a cascading threshold which is the same as the bifurcation point. When systems are coupled, this is no longer true and the cascading threshold is different from the bifurcation point of the fixed point solutions. This change in the cascading threshold caused by the interdependence of the system can have an impact on the "safe operation" of interdependent infrastructure systems by changing the critical point and even the power law exponent.

The intensity of single-crystal Bragg peaks obtained by mapping neutron time-of-flight event data into reciprocal space and integrating in various ways is compared. These methods include spherical integration with a fixed radius, ellipsoid fitting and integration of the peak intensity, and one-dimensional peak profile fitting. In comparison to intensities obtained by integrating in real detector histogram space, the data integrated in reciprocal space result in better agreement factors and more accurate atomic parameters. Furthermore, structure refinement using integrated intensities from one-dimensional profile fitting is demonstrated to be more accurate than simple peak-minus-background integration.

This presentation will review developments on the integration of advanced modeling and simulation techniques into the analysis step of experimental data obtained at the Spallation Neutron Source. A workflow framework for the purpose of refining molecular mechanics forcefields against quasi-elastic neutron scattering data is presented. The workflow combines software components to submit model simulations to remote high performance computers, a message broker interface for communications between the optimizer engine and the simulation production step, and tools to convolve the simulated data with the experimental resolution. A test application shows the correction to a popular fixed-charge water model in order to account polarization effects due to the presence of solvated ions. Future enhancements to the refinement workflow are discussed. This work is funded through the DOE Center for Accelerating Materials Modeling.