Modern High-performance Computing Research Articles

Accelerating radio astronomy imaging with RICK

This paper presents an implementation of radio astronomy imaging algorithms on modern High Performance Computing (HPC) infrastructures, exploiting distributed memory parallelism and acceleration throughout multiple GPUs. Our code, called RICK (Radio Imaging Code Kernels), is capable of performing the major steps of the w-stacking algorithm presented in Offringa et al. (2014) both inter- and intra-node, and in particular has the possibility to run entirely on the GPU memory, minimising the number of data transfers between CPU and GPU. This feature, especially among multiple GPUs, is critical given the huge sizes of radio datasets involved.After a detailed description of the new implementations of the code with respect to the first version presented in Gheller et al. (2023), we analyse the performances of the code for each step involved in its execution. We also discuss the pros and cons related to an accelerated approach to this problem and its impact on the overall behaviour of the code. Such approach to the problem results in a significant improvement in terms of runtime with respect to the CPU version of the code, as long as the amount of computational resources does not exceed the one requested by the size of the problem: the code, in fact, is now limited by the communication costs, with the computation that gets heavily reduced by the capabilities of the accelerators.

Network-on-Chip (NoC) Applications for IoT-Enabled Chip Systems: Latest Designs and Modern Applications

Network-on-Chip (NoC) technology has emerged as a critical innovation in the field of integrated circuit design, addressing the growing demands for efficient, scalable, and high-performance on-chip communication. This review provides a comprehensive examination of NoC, highlighting its architecture, key features, and current applications. NoC leverages a network-based approach, utilizing routers and links to interconnect various components on a chip, thereby overcoming the limitations of traditional System-on-chip (SoC) architectures that rely on shared bus systems. The review underscores NoC’s superior communication efficiency, scalability, and reduced latency, which are essential for modern high-performance computing, multi-core processors, and advanced embedded systems. It also explores NoC’s ability to optimize power consumption through dynamic voltage scaling and power gating, making it a favorable choice for energy-efficient designs. Despite its complexity, NoC’s modularity and adaptability make it suitable for a wide range of applications, from artificial intelligence accelerators to high-performance data centers. In conclusion, NoC stands as a transformative technology that enhances the performance and scalability of integrated circuits, paving the way for more sophisticated and powerful electronic systems. As the demand for higher computational capabilities and efficiency continues to rise, NoC is poised to play an increasingly vital role in the advancement of electronic design and architecture.

Very-Large-Scale GPU-Accelerated Nuclear Gradient of Time-Dependent Density Functional Theory with Tamm-Dancoff Approximation and Range-Separated Hybrid Functionals.

Modern graphics processing units (GPUs) provide an unprecedented level of computing power. In this study, we present a high-performance, multi-GPU implementation of the analytical nuclear gradient for Kohn-Sham time-dependent density functional theory (TDDFT), employing the Tamm-Dancoff approximation (TDA) and Gaussian-type atomic orbitals as basis functions. We discuss GPU-efficient algorithms for the derivatives of electron repulsion integrals and exchange-correlation functionals within the range-separated scheme. As an illustrative example, we calculate the TDA-TDDFT gradient of the S1 state of a full-scale green fluorescent protein with explicit water solvent molecules, totaling 4353 atoms, at the ωB97X/def2-SVP level of theory. Our algorithm demonstrates favorable parallel efficiencies on a high-speed distributed system equipped with 256 Nvidia A100 GPUs, achieving >70% with up to 64 GPUs and 31% with 256 GPUs, effectively leveraging the capabilities of modern high-performance computing systems.

The Advanced Computing Hub at BSC: improving fusion codes following modern software engineering standards

Several dedicated High-Performance Computing (HPC) centers provide essential expertise and support in developing a suitable portfolio of EUROfusion standard codes. Barcelona Supercomputing Center (BSC) is one of these HPC hubs involved in this complex task. Several fusion codes were selected, installed and analyzed to meet the developers’ requirements, ranging from portability to GPU, improving the performance, getting better data management, extending the capacity of coupling with other codes, etc. In this paper, we will describe the work developed by BSC and some of the tasks carried out in this project. We will explain briefly how the project is faced and the work required to create good quality codes, i.e. robust and trustable software capable of running efficiently in modern HPC systems.

Implementation and evaluation of diabatic advection in the Lagrangian transport model MPTRAC 2.6

Abstract. Diabatic transport schemes with hybrid zeta coordinates, which follow isentropes in the stratosphere, are known to greatly improve Lagrangian transport calculations compared to the kinematic approach. However, some Lagrangian transport calculations with a diabatic approach, such as the Chemical Lagrangian Transport Model of the Stratosphere (CLaMS), are not well prepared to run on modern high-performance computing (HPC) architectures. Here, we implemented and evaluated a new diabatic transport scheme in the Massive-Parallel Trajectory Calculations (MPTRAC) model. While MPTRAC can be used either with shared-memory multiprocessing on CPUs or with GPUs to offload computationally intensive calculations, making it flexible for many HPC applications, it has been limited to kinematic trajectories in pressure coordinates. The extended modelling approach now enables the use of either kinematic or diabatic vertical velocities and the coupling of different MPTRAC modules based on pressure or hybrid zeta coordinates. This study focus on the accuracy of the implementation in comparison to the CLaMS model. The evaluation of the new transport scheme in MPTRAC shows that, after 90 d of forward calculations, distributions of air parcels in the upper troposphere and lower stratosphere (UTLS) are almost identical for MPTRAC and CLaMS. No significant bias between the two Lagrangian models was found. Furthermore, after 1 d, internal uncertainties (e.g. due to interpolation or the numerical integration method) in the Lagrangian transport calculations are at least 1 order of magnitude smaller than external uncertainties (e.g. from reanalysis selection or downsampling of ERA5). Differences between trajectories using either CLaMS or MPTRAC are on the order of the combined internal uncertainties within MPTRAC. Since the largest systematic differences are caused by the reanalysis and the vertical velocity (diabatic vs. kinematic), the results support the development efforts for trajectory codes that can access the full resolution of ERA5 in combination with diabatic vertical velocities. This work is part of a larger effort to adapt Lagrangian transport in state-of-the-art models such as CLaMS and MPTRAC to current and future HPC architectures and exascale applications.

On the feasibility of overnight industrial high-fidelity simulations of CSP technologies on modern HPC systems

In the last decades, computational fluid dynamics (CFD) has become a standard design tool in many fields, such as the automotive, aeronautical, and renewable energy industries. The driving force behind this is the development of numerical techniques in conjunction with the progress of high-performance computing (HPC) systems. However, simulation time remains the most limiting factor for large-eddy simulations (LES) to be adopted in the industry. A consensus exists that, to be feasible, LES simulations should be completed overnight In this context, this work assesses the feasibility of overnight LES simulations on GPU-accelerated supercomputers with TFA, our novel in-house code, which relies on a symmetry-preserving discretisation for unstructured collocated grids that, apart from being virtually free of artificial dissipation, is shown to be unconditionally stable. The study cases will be taken from central receivers used in concentrated solar power (CSP) plants, and a comparison with open-source CFD codes will be made.

Swift : a modern highly parallel gravity and smoothed particle hydrodynamics solver for astrophysical and cosmological applications

ABSTRACT Numerical simulations have become one of the key tools used by theorists in all the fields of astrophysics and cosmology. The development of modern tools that target the largest existing computing systems and exploit state-of-the-art numerical methods and algorithms is thus crucial. In this paper, we introduce the fully open-source highly-parallel, versatile, and modular coupled hydrodynamics, gravity, cosmology, and galaxy-formation code Swift. The software package exploits hybrid shared- and distributed-memory task-based parallelism, asynchronous communications, and domain-decomposition algorithms based on balancing the workload, rather than the data, to efficiently exploit modern high-performance computing cluster architectures. Gravity is solved for using a fast-multipole-method, optionally coupled to a particle mesh solver in Fourier space to handle periodic volumes. For gas evolution, multiple modern flavours of Smoothed Particle Hydrodynamics are implemented. Swift also evolves neutrinos using a state-of-the-art particle-based method. Two complementary networks of sub-grid models for galaxy formation as well as extensions to simulate planetary physics are also released as part of the code. An extensive set of output options, including snapshots, light-cones, power spectra, and a coupling to structure finders are also included. We describe the overall code architecture, summarize the consistency and accuracy tests that were performed, and demonstrate the excellent weak-scaling performance of the code using a representative cosmological hydrodynamical problem with ≈300 billion particles. The code is released to the community alongside extensive documentation for both users and developers, a large selection of example test problems, and a suite of tools to aid in the analysis of large simulations run with Swift.

FLEW: A DNS Solver for Compressible Flows in Generalized Curvilinear Coordinates

We present FLEW, an in-house high-fidelity solver for direct numerical simulation (DNS) of turbulent compressible flows over arbitrary shaped geometries. FLEW solves the Navier–Stokes equations written in a generalized curvilinear coordinate system, in which the surface coordinates are non-orthogonal, whereas the third axis is normal to the surface. Spatial discretization relies on high-order finite-difference schemes. The convective terms are discretized using an hybrid approach, combining the near-zero numerical dissipation provided by central approximations with the robustness of weighted essentially non-oscillatory (WENO) schemes, required to capture shock waves. Central schemes are stabilized using a skew-symmetric-like splitting of convective derivatives, endowing the solver with the energy-preserving property in the inviscid limit. The maximum order of accuracy is eighth for central schemes (also used for viscous terms discretization) and seventh for WENO. The code is oriented to modern high-performance computing (HPC) platforms thanks to message passing interface (MPI) parallelization and the ability to run on graphics processing unit (GPU) architectures. Reliability, accuracy and robustness of the code are assessed in the low-subsonic, transonic and supersonic regimes. We present the results of several benchmarks, namely the inviscid Taylor–Green vortex, turbulent curved channel flow, transonic laminar flow over a NACA 0012 airfoil and turbulent supersonic ramp flow. The results for all configurations proved to be in excellent agreement with previous studies.

ControlPULP: A RISC-V On-Chip Parallel Power Controller for Many-Core HPC Processors with FPGA-Based Hardware-In-The-Loop Power and Thermal Emulation

High-performance computing (HPC) processors are nowadays integrated cyber-physical systems demanding complex and high-bandwidth closed-loop power and thermal control strategies. To efficiently satisfy real-time multi-input multi-output (MIMO) optimal power requirements, high-end processors integrate an on-die power controller system (PCS). While traditional PCSs are based on a simple microcontroller (MCU)-class core, more scalable and flexible PCS architectures are required to support advanced MIMO control algorithms for managing the ever-increasing number of cores, power states, and process, voltage, and temperature variability. This paper presents ControlPULP, an open-source, HW/SW RISC-V parallel PCS platform consisting of a single-core MCU with fast interrupt handling coupled with a scalable multi-core programmable cluster accelerator and a specialized DMA engine for the parallel acceleration of real-time power management policies. ControlPULP relies on FreeRTOS to schedule a reactive power control firmware (PCF) application layer. We demonstrate ControlPULP in a power management use-case targeting a next-generation 72-core HPC processor. We first show that the multi-core cluster accelerates the PCF, achieving 4.9x speedup compared to single-core execution, enabling more advanced power management algorithms within the control hyper-period at a shallow area overhead, about 0.1% the area of a modern HPC CPU die. We then assess the PCS and PCF by designing an FPGA-based, closed-loop emulation framework that leverages the heterogeneous SoCs paradigm, achieving DVFS tracking with a mean deviation within 3% the plant’s thermal design power (TDP) against a software-equivalent model-in-the-loop approach. Finally, we show that the proposed PCF compares favorably with an industry-grade control algorithm under computational-intensive workloads.

Comparison of methods for optimizing the speed of reading/writing drives

The objects of this study are data storage devices of various types and levels of complexity, as well as the principles of their operation. They are complex technical systems that include many components and are characterized by a high degree of integration. The subject of the research is to study the main characteristics of hard drives and solid-state drives. Their structure, functional features, principles of operation and ways of optimization are important. The purpose of the study is to determine the most effective methods for optimizing the operation of these devices. This includes aspects such as memory management, load balancing, power management, and others. The results of this research can be used to improve data efficiency, improve the performance of data storage systems and create new technologies in this area. This study examines the performance of various disk storage solutions through a series of tests aimed at understanding speed and dependence on external factors. The main conclusions of the study reflect the importance of the integrated use of optimization approaches to improve the speed of reading and writing data. Optimizing the processes of reading and writing data is critically important for modern high-performance computing systems, as well as for applications that require quick access to large amounts of information. The improved techniques used in the course of the study contribute to a significant increase in the performance of data storage devices. They take into account the specifics of various types of storage devices, including hard drives and solid-state drives, and offer optimization approaches that take into account their unique characteristics. Overall, the results of this study provide valuable insights into the principles of optimizing data storage, and they can serve as a basis for developing new strategies and solutions in this important area of information technology. This study represents a significant contribution to the scientific understanding of optimizing data reading and writing processes, and its findings may have long-term implications for the development of data storage technologies.

Distributed Multi-GPU Ab Initio Density Matrix Renormalization Group Algorithm with Applications to the P-Cluster of Nitrogenase.

The presence of many degenerate d/f orbitals makes polynuclear transition-metal compounds, such as iron-sulfur clusters in nitrogenase, challenging for state-of-the-art quantum chemistry methods. To address this challenge, we present the first distributed multi-graphics processing unit (GPU) ab initio density matrix renormalization group (DMRG) algorithm suitable for modern high-performance computing (HPC) infrastructures. The central idea is to parallelize the most computationally intensive part─the multiplication of O(K2) operators with a trial wave function, where K is the number of spatial orbitals, by combining operator parallelism for distributing the workload with a batched algorithm for performing contractions on GPU. With this new implementation, we are able to reach an unprecedentedly large bond dimension D = 14,000 on 48 GPUs (NVIDIA A100 80 GB SXM) for an active space model (114 electrons in 73 active orbitals) of the P-cluster, which is nearly 3 times larger than the bond dimensions reported in previous DMRG calculations for the same system using only central processing units (CPUs).

Streamlined Algorithms for Large-Scale Matrix Computations in High-Performance Computing

In the realm of high-performance computing (HPC), large-scale matrix computations are fundamental to a myriad of scientific and engineering applications, from simulations in physics and engineering to data analysis in machine learning. This paper addresses the critical need for efficient and scalable algorithms capable of handling the growing computational demands of these applications. This study explores advanced algorithmic strategies that leverage the architecture of modern HPC systems, including parallelism, distributed computing, and optimized memory usage. The proposed algorithms are designed to minimize computational complexity and maximize performance, ensuring that largescale matrix operations can be executed swiftly and accurately. One of the key contributions of this work is the development of parallel algorithms that distribute the computational workload across multiple processors, significantly reducing execution time. By employing techniques such as data partitioning and workload balancing, the algorithms ensure optimal use of available computational resources, thus enhancing scalability. Furthermore, the paper introduces innovative approaches to memory management, which are crucial for handling large matrices that exceed the memory capacity of individual computing nodes. These approaches include efficient data storage formats and dynamic memory allocation strategies that reduce overhead and improve data access speeds. Another critical aspect of this research is the application of these algorithms to real-world problems. The paper presents case studies in fields such as climate modeling, structural analysis, and machine learning, demonstrating the practical benefits and performance gains achieved through the use of streamlined matrix computation algorithms. The results show substantial improvements in computation time and resource utilization, validating the effectiveness of the proposed methods. Additionally, the paper discusses the integration of these algorithms into existing HPC frameworks and libraries, making them accessible to a broader community of researchers and practitioners. By providing detailed implementation guidelines and performance benchmarks, the study ensures that the benefits of these advanced algorithms can be readily harnessed in various applications. The research also delves into the theoretical underpinnings of the proposed algorithms, offering a rigorous analysis of their computational complexity and scalability. In conclusion, this paper makes significant strides in addressing the challenges of large-scale matrix operations. Through the development of efficient, scalable, and practical algorithms, this research enhances the capability of HPC systems to tackle complex computational tasks, paving the way for advancements in various scientific and engineering domains. The proposed methods not only improve performance and resource utilization but also contribute to the broader goal of making high-performance computing more accessible and effective for large-scale data analysis and simulation.

Streamlined Algorithms for Large-Scale Matrix Computations in High-Performance Computing

In the realm of high-performance computing (HPC), large-scale matrix computations are fundamental to a myriad of scientific and engineering applications, from simulations in physics and engineering to data analysis in machine learning. This paper addresses the critical need for efficient and scalable algorithms capable of handling the growing computational demands of these applications. This study explores advanced algorithmic strategies that leverage the architecture of modern HPC systems, including parallelism, distributed computing, and optimized memory usage. The proposed algorithms are designed to minimize computational complexity and maximize performance, ensuring that largescale matrix operations can be executed swiftly and accurately. One of the key contributions of this work is the development of parallel algorithms that distribute the computational workload across multiple processors, significantly reducing execution time. By employing techniques such as data partitioning and workload balancing, the algorithms ensure optimal use of available computational resources, thus enhancing scalability. Furthermore, the paper introduces innovative approaches to memory management, which are crucial for handling large matrices that exceed the memory capacity of individual computing nodes. These approaches include efficient data storage formats and dynamic memory allocation strategies that reduce overhead and improve data access speeds. Another critical aspect of this research is the application of these algorithms to real-world problems. The paper presents case studies in fields such as climate modeling, structural analysis, and machine learning, demonstrating the practical benefits and performance gains achieved through the use of streamlined matrix computation algorithms. The results show substantial improvements in computation time and resource utilization, validating the effectiveness of the proposed methods. Additionally, the paper discusses the integration of these algorithms into existing HPC frameworks and libraries, making them accessible to a broader community of researchers and practitioners. By providing detailed implementation guidelines and performance benchmarks, the study ensures that the benefits of these advanced algorithms can be readily harnessed in various applications. The research also delves into the theoretical underpinnings of the proposed algorithms, offering a rigorous analysis of their computational complexity and scalability. In conclusion, this paper makes significant strides in addressing the challenges of large-scale matrix operations. Through the development of efficient, scalable, and practical algorithms, this research enhances the capability of HPC systems to tackle complex computational tasks, paving the way for advancements in various scientific and engineering domains. The proposed methods not only improve performance and resource utilization but also contribute to the broader goal of making high-performance computing more accessible and effective for large-scale data analysis and simulation.

Predicting Software Defects in Hybrid MPI and OpenMP Parallel Programs Using Machine Learning

High-performance computing (HPC) and its supercomputers are essential for solving the most difficult issues in many scientific computing domains. The proliferation of computational resources utilized by HPC systems has resulted in an increase in the associated error rates. As such, modern HPC systems promote a hybrid programming style that integrates the message-passing interface (MPI) and open multi-processing (OpenMP). However, this integration often leads to complex defects, such as deadlocks and race conditions, that are challenging to detect and resolve. This paper presents a novel approach: using machine learning algorithms to predict defects in C++-based systems by employing hybrid MPI and OpenMP models. We focus on employing a balanced dataset to enhance prediction accuracy and reliability. Our study highlights the effectiveness of the support vector machine (SVM) classifier, enhanced with term frequency (TF) and recursive feature elimination (RFE) techniques, which demonstrates superior accuracy and performance in defect prediction when compared to other classifiers. This research contributes significantly to the field by providing a robust method for early defect detection in hybrid programming environments, thereby reducing development time, costs and improving the overall reliability of HPC systems.

PARALiA: A Performance Aware Runtime for Auto-tuning Linear Algebra on Heterogeneous Systems

Dense linear algebra operations appear very frequently in high-performance computing (HPC) applications, rendering their performance crucial to achieve optimal scalability. As many modern HPC clusters contain multi-GPU nodes, BLAS operations are frequently offloaded on GPUs, necessitating the use of optimized libraries to ensure good performance. Unfortunately, multi-GPU systems are accompanied by two significant optimization challenges: data transfer bottlenecks as well as problem splitting and scheduling in multiple workers (GPUs) with distinct memories. We demonstrate that the current multi-GPU BLAS methods for tackling these challenges target very specific problem and data characteristics, resulting in serious performance degradation for any slightly deviating workload. Additionally, an even more critical decision is omitted because it cannot be addressed using current scheduler-based approaches: the determination of which devices should be used for a certain routine invocation. To address these issues we propose a model-based approach: using performance estimation to provide problem-specific autotuning during runtime. We integrate this autotuning into an end-to-end BLAS framework named PARALiA. This framework couples autotuning with an optimized task scheduler, leading to near-optimal data distribution and performance-aware resource utilization. We evaluate PARALiA in an HPC testbed with 8 NVIDIA-V100 GPUs, improving the average performance of GEMM by 1.7× and energy efficiency by 2.5× over the state-of-the-art in a large and diverse dataset and demonstrating the adaptability of our performance-aware approach to future heterogeneous systems.

OBTAINING BULK PRODUCTS FROM CU-SIC METAL-MATRIX COMPOSITE FOR ENERGY-EFFICIENT HEAT-CONDUCTING SYSTEMS

Link for citation: Nikitin D.S., Shanenkov I.I., Nassyrbayev A., Vympina Yu. N., Orlova E.G., Ivashutenko A.S., Sivkov A.A. Obtaining bulk products from Cu-SiC metal-matrix composite for energy-efficient heat-conducting systems. Bulletin of the Tomsk Polytechnic University. Geo Аssets Engineering, 2023, vol. 334, no. 7, рр. 93-101. In Rus. The relevance of the research is associated with the rapid development of modern high-performance computing systems, superneurocomputers and artificial intelligence devices. Today such development is held back largely due to the lack of an effective cooling system for high-power elements of their structures. Composite materials Cu-SiC with improved physical, mechanical and thermophysical characteristics can be used to solve problems of heat removal intensification. The main aim of the research is to obtain bulk products from metal-matrix composite Cu-10% SiC with improved physical, mechanical and thermal characteristics by spark plasma sintering. Objects of the research are bulk products from metal-matrix composite Cu-10%SiC. The samples were obtained by spark plasma sintering at temperatures of 700, 750, 800, 850 °C and a pressure of 60 MPa. Methods: spark plasma sintering, X-ray diffractometry (X-ray phase analysis), scanning electron microscopy, indentation, laser flash method. Results. Experimental studies have been carried out to obtain bulk metal-matrix composites with a copper matrix and the addition of reinforcing superhard particles of silicon carbide Cu-10%SiC. The compaction of dispersed composite materials was carried out by spark plasma sintering at various temperatures of 700, 750, 800, and 850 °C. The microstructure and composition of initial dispersed materials and final bulk products have been studied. It is shown that the spark plasma sintering method has advantages for obtaining relatively dense materials with high physical, mechanical and thermal properties. Analysis of the obtained samples showed the formation of a dense (up to ~88 %) homogeneous fine-grained composite structure. The greatest densification of the material is achieved at the highest sintering temperature of 850 °C, which causes this sample to demonstrate the maximum hardness (H=3,63 GPa) and Young's modulus (E=159,63 GPa), as well as the thermal conductivity at room temperature (λ=223 W/m K). The obtained composite materials can be used as structural and functional materials for energy-efficient heat-conducting systems.

Mixed precision support in HPC applications: What about reliability?

In their quest for exascale and beyond, High-Performance Computing (HPC) systems continue becoming ever larger and more complex. Application developers, on the other hand, leverage novel methods to improve the efficiency of their own codes: a recent trend is the use of floating-point mixed precision, or the careful interlocking of single- and double-precision arithmetic, as a tool to improve performance as well as reduce network and memory boundedness. However, while it is known that modern HPC systems suffer hardware faults at daily rates, the impact of reduced precision on application reliability is yet to be explored. In this work we aim to fill this gap: first, we propose a qualitative survey to identify the branches of HPC where mixed precision is most popular. Second, we show the results of instruction-level fault injection experiments on a variety of representative HPC workloads, comparing vulnerability to Silent Data Errors (SDEs) under different numerical configurations. Our experiments indicate that use of single and mixed precision leads to comparatively more frequent and more severe SDEs, with concerning implications regarding their use on extreme-scale, fault-prone HPC platforms.

A parallel, distributed memory implementation of the adaptive sampling configuration interaction method.

The many-body simulation of quantum systems is an active field of research that involves several different methods targeting various computing platforms. Many methods commonly employed, particularly coupled cluster methods, have been adapted to leverage the latest advances in modern high-performance computing. Selected configuration interaction (sCI) methods have seen extensive usage and development in recent years. However, the development of sCI methods targeting massively parallel resources has been explored only in a few research works. Here, we present a parallel, distributed memory implementation of the adaptive sampling configuration interaction approach (ASCI) for sCI. In particular, we will address the key concerns pertaining to the parallelization of the determinant search and selection, Hamiltonian formation, and the variational eigenvalue calculation for the ASCI method. Load balancing in the search step is achieved through the application of memory-efficient determinant constraints originally developed for the ASCI-PT2 method. The presented benchmarks demonstrate near optimal speedup for ASCI calculations of Cr2 (24e, 30o) with 106, 107, and 3 × 108 variational determinants on up to 16 384 CPUs. To the best of the authors' knowledge, this is the largest variational ASCI calculation to date.

Energy-Efficient Parallel Computing: Challenges to Scaling

The energy consumption of Information and Communications Technology (ICT) presents a new grand technological challenge. The two main approaches to tackle the challenge include the development of energy-efficient hardware and software. The development of energy-efficient software employing application-level energy optimization techniques has become an important category owing to the paradigm shift in the composition of digital platforms from single-core processors to heterogeneous platforms integrating multicore CPUs and graphics processing units (GPUs). In this work, we present an overview of application-level bi-objective optimization methods for energy and performance that address two fundamental challenges, non-linearity and heterogeneity, inherent in modern high-performance computing (HPC) platforms. Applying the methods requires energy profiles of the application’s computational kernels executing on the different compute devices of the HPC platform. Therefore, we summarize the research innovations in the three mainstream component-level energy measurement methods and present their accuracy and performance tradeoffs. Finally, scaling the optimization methods for energy and performance is crucial to achieving energy efficiency objectives and meeting quality-of-service requirements in modern HPC platforms and cloud computing infrastructures. We introduce the building blocks needed to achieve this scaling and conclude with the challenges to scaling. Briefly, two significant challenges are described, namely fast optimization methods and accurate component-level energy runtime measurements, especially for components running on accelerators.

HDF5eis: A storage and input/output solution for big multidimensional time series data from environmental sensors

Modern high-performance computing (HPC) tasks overwhelm conventional geophysical data formats. We describe a new data schema called HDF5eis (read H-D-F-size) for handling big multidimensional time series data from environmental sensors in HPC applications and implement a freely available Python application programming interface (API) for building and processing HDF5eis files. HDF5eis augments the popular Hierarchical Data Format 5 with a minimal set of additional conventions that facilitate fast and flexible data input and output protocols for regularly sampled (in time) data with any number of dimensions. HDF5eis supports arbitrary ancillary data (e.g., metadata) storage in columnar format or as UTF-8 encoded byte streams alongside time series data. Our HDF5eis API enables simple and efficient access to big data sets distributed across a potentially large number of small heterogeneous files through a single point of access. HDF5eis outperforms conventional seismic data formats by up to two orders of magnitude in terms of random read access times. We contribute HDF5eis as an operational tool and an experimental draft proposal that will help establish the next generation of data standards in the earth sciences.