Search

Neural networks (NNs) have proven to be a viable alternative to traditional direct numerical algorithms, with the potential to accelerate computational time by several orders of magnitude. In the present paper we study the use of encoder-decoder convolutional neural network (CNN) as surrogates for steady-state diffusion solvers. The construction of such surrogates requires the selection of an appropriate task, network architecture, training set structure and size, loss function, and training algorithm hyperparameters. It is well known that each of these factors can have a significant impact on the performance of the resultant model. Our approach employs an encoder-decoder CNN architecture, which we posit is particularly well-suited for this task due to its ability to effectively transform data, as opposed to merely compressing it. We systematically evaluate a range of loss functions, hyperparameters, and training set sizes. Our results indicate that increasing the size of the training set has a substantial effect on reducing performance fluctuations and overall error. Additionally, we observe that the performance of the model exhibits a logarithmic dependence on the training set size. Furthermore, we investigate the effect on model performance by using different subsets of data with varying features. Our results highlight the importance of sampling the configurational space in an optimal manner, as this can have a significant impact on the performance of the model and the required training time. In conclusion, our results suggest that training a model with a pre-determined error performance bound is not a viable approach, as it does not guarantee that edge cases with errors larger than the bound do not exist. Furthermore, as most surrogate tasks involve a high dimensional landscape, an ever increasing training set size is, in principle, needed, however it is not a practical solution., Comment: 21 ps, 17 figs, 8 ts

The Data Science domain has expanded monumentally in both research and industry communities during the past decade, predominantly owing to the Big Data revolution. Artificial Intelligence (AI) and Machine Learning (ML) are bringing more complexities to data engineering applications, which are now integrated into data processing pipelines to process terabytes of data. Typically, a significant amount of time is spent on data preprocessing in these pipelines, and hence improving its e fficiency directly impacts the overall pipeline performance. The community has recently embraced the concept of Dataframes as the de-facto data structure for data representation and manipulation. However, the most widely used serial Dataframes today (R, pandas) experience performance limitations while working on even moderately large data sets. We believe that there is plenty of room for improvement by taking a look at this problem from a high-performance computing point of view. In a prior publication, we presented a set of parallel processing patterns for distributed dataframe operators and the reference runtime implementation, Cylon [1]. In this paper, we are expanding on the initial concept by introducing a cost model for evaluating the said patterns. Furthermore, we evaluate the performance of Cylon on the ORNL Summit supercomputer.

The data engineering and data science community has embraced the idea of using Python & R dataframes for regular applications. Driven by the big data revolution and artificial intelligence, these applications are now essential in order to process terabytes of data. They can easily exceed the capabilities of a single machine, but also demand significant developer time & effort. Therefore it is essential to design scalable dataframe solutions. There have been multiple attempts to tackle this problem, the most notable being the dataframe systems developed using distributed computing environments such as Dask and Ray. Even though Dask/Ray distributed computing features look very promising, we perceive that the Dask Dataframes/Ray Datasets still have room for optimization. In this paper, we present CylonFlow, an alternative distributed dataframe execution methodology that enables state-of-the-art performance and scalability on the same Dask/Ray infrastructure (thereby supercharging them!). To achieve this, we integrate a high performance dataframe system Cylon, which was originally based on an entirely different execution paradigm, into Dask and Ray. Our experiments show that on a pipeline of dataframe operators, CylonFlow achieves 30x more distributed performance than Dask Dataframes. Interestingly, it also enables superior sequential performance due to the native C++ execution of Cylon. We believe the success of Cylon & CylonFlow extends beyond the data engineering domain, and can be used to consolidate high performance computing and distributed computing ecosystems.

Data pre-processing is a fundamental component in any data-driven application. With the increasing complexity of data processing operations and volume of data, Cylon, a distributed dataframe system, is developed to facilitate data processing both as a standalone application and as a library, especially for Python applications. While Cylon shows promising performance results, we experienced difficulties trying to integrate with frameworks incompatible with the traditional Message Passing Interface (MPI). While MPI implementations encompass scalable and efficient communication routines, their process launching mechanisms work well with mainstream HPC systems but are incompatible with some environments that adopt their own resource management systems. In this work, we alleviated this issue by directly integrating the Unified Communication X (UCX) framework, which supports a variety of classic HPC and non-HPC process-bootstrapping mechanisms as our communication framework. While we experimented with our methodology on Cylon, the same technique can be used to bring MPI communication to other applications that do not employ MPI's built-in process management approach.

A foundational set of findable, accessible, interoperable, and reusable (FAIR) principles were proposed in 2016 as prerequisites for proper data management and stewardship, with the goal of enabling the reusability of scholarly data. The principles were also meant to apply to other digital assets, at a high level, and over time, the FAIR guiding principles have been re-interpreted or extended to include the software, tools, algorithms, and workflows that produce data. FAIR principles are now being adapted in the context of AI models and datasets. Here, we present the perspectives, vision, and experiences of researchers from different countries, disciplines, and backgrounds who are leading the definition and adoption of FAIR principles in their communities of practice, and discuss outcomes that may result from pursuing and incentivizing FAIR AI research. The material for this report builds on the FAIR for AI Workshop held at Argonne National Laboratory on June 7, 2022., Comment: 10 pages, comments welcome!

PI: G. Fox, Co-PIs: Madhav Marathe, Shantenu Jha, Judy Qiu, Fusheng Wang Institutions: Arizona State, Indiana (lead), Kansas, Rutgers, Stony Brook, Virginia, and Utah.NSF 1443054 “Middleware and High-Performance Analytics Libraries for Scalable Data Science” is a collaboration between 7 collaborating Universities at Arizona State, Indiana (lead), Kansas, Rutgers, Stony Brook, Virginia, and Utah. It addresses the intersection of HPC and Big Data computing with several different application areas or communities driving the requirements for software systems and algorithms. The base architecture includes the HPC-ABDS, High-Performance Computing Enhanced Apache Big Data Stack, and application use cases identifying key features that determine software and algorithm requirements. The middleware includes the Harp-DAAL collective communication layer, Twister2 Big Data toolkit, and RADICAL pilot jobs for batch and streaming applications. The SPIDAL Scalable Parallel Interoperable Data Analytics Library includes core machine-learning, image processing, and the application communities, Network science, Polar Science, Biomolecular Simulations, Pathology, and Spatial systems. Recent work focuses on the integration of ML with HPC in HPCafterML in biomolecular simulations and a broad study of HPCforML.

Performance analysis is a daunting job, especially for the rapid-evolving accelerator technologies. The Roofline Scaling Trajectories technique aims at diagnosing various performance bottlenecks for GPU programming models through the visually intuitive Roofline plots. In this work, we introduce the use of the Roofline Scaling Trajectories to capture major performance bottlenecks on NVIDIA Volta GPU architectures, such as warp efficiency, occupancy, and locality. Using this analysis technique, we explain the performance characteristics of the NAS Parallel Benchmarks (NPB) written with two programming models, CUDAand OpenACC. We present the influence of the programming model on the performance and scaling characteristics. We also leverage the insights of the Roofline Scaling Trajectory analysis to tune some of the NAS Parallel Benchmarks, achieving upto 2$$\times $$ speedup.

In this paper, we describe the efforts in taking advantage of the latest IT developments, (especially of the development of mobile computing devices such as tablets or phones) to create a comprehensive architecture for face-to-face meetings support and groups evolution analysis. The main purpose of the architecture is to mitigate frequent meetings problems by providing state-of-the-art technological support to groups or teams. This paper introduces ReflectWorld, the distributed architecture created on top of the principles of Reflect Table, a meeting support and analysis system centered on individuals’ participation and interactions.