Search NASA⌕ Search

Engineering topics

Deslippe, Jack

Publications and source records attributed to Deslippe, Jack.

At least 37 records · Page 2

Preparing NERSC users for Cori, a Cray XC40 system with Intel many integrated cores

The newest NERSC supercomputer Cori is a Cray XC40 system consisting of 2,388 Intel Xeon Haswell nodes and 9,688 Intel Xeon-Phi “Knights Landing” (KNL) nodes. Compared to the Xeon-based clusters NERSC users are familiar with, optimal performance on Cori requires consideration of KNL mode settings; process, thread, and memory affinity; fine-grain parallelization; vectorization; and use of the high-bandwidth MCDRAM memory. This paper describes our efforts preparing NERSC users for KNL through the NERSC Exascale Science Application Program, Web documentation, and user training. We discuss how we configured the Cori system for usability and productivity, addressing programming concerns, batch system configurations, and default KNL cluster and memory modes. Here, system usage data, job completion analysis, programming and running jobs issues, and a few successful user stories on KNL are presented.

97 MATHEMATICS AND COMPUTING↗

Fusion PIC code performance analysis on the Cori KNL system

We study the attainable performance of Particle-In-Cell codes on the Cori KNL system by analyzing a miniature particle push application based on the fusion PIC code XGC1. We start from the most basic building blocks of a PIC code and build up the complexity to identify the kernels that cost the most in performance and focus optimization efforts there. Particle push kernels operate at high AI and are not likely to be memory bandwidth or even cache bandwidth bound on KNL. Therefore, we see only minor benefits from the high bandwidth memory available on KNL, and achieving good vectorization is shown to be the most beneficial optimization path with theoretical yield of up to 8x speedup on KNL. In practice we are able to obtain up to a 4x gain from vectorization due to limitations set by the data layout and memory latency.

97 MATHEMATICS AND COMPUTING↗

Roofline Analysis in the Intel® Advisor to Deliver Optimized Performance for applications on Intel® Xeon Phi™ Processor

In this session we show, in two case studies, how the roofline feature of Intel Advisor has been utilized to optimize the performance of kernels of the XGC1 and PICSAR codes in preparation for Intel Knights Landing architecture. The impact of the implemented optimizations and the benefits of using the automatic roofline feature of Intel Advisor to study performance of large applications will be presented. This demonstrates an effective optimization strategy that has enabled these science applications to achieve up to 4.6 times speed-up and prepare for future exascale architectures. # Goal/Relevance of Session The roofline model [1,2] is a powerful tool for analyzing the performance of applications with respect to the theoretical peak achievable on a given computer architecture. It allows one to graphically represent the performance of an application in terms of operational intensity, i.e. the ratio of flops performed and bytes moved from memory in order to guide optimization efforts. Given the scale and complexity of modern science applications, it can often be a tedious task for the user to perform the analysis on the level of functions or loops to identify where performance gains can be made. With new Intel tools, it is now possible to automate this task, as well as base the estimates of peak performance on measurements rather than vendor specifications. The goal of this session is to demonstrate how the roofline feature of Intel Advisor can be used to balance memory vs. computation related optimization efforts and effectively identify performance bottlenecks. A series of typical optimization techniques: cache blocking, structure refactoring, data alignment, and vectorization illustrated by the kernel cases will be addressed. # Description of the codes ## XGC1 The XGC1 code [3] is a magnetic fusion Particle-In-Cell code that uses an unstructured mesh for its Poisson solver that allows it to accurately resolve the edge plasma of a magnetic fusion device. After recent optimizations to its collision kernel [4], most of the computing time is spent in the electron push (pushe) kernel, where these optimization efforts have been focused. The kernel code scaled well with MPI+OpenMP but had almost no automatic compiler vectorization, in part due to indirect memory addresses and in part due to low trip counts of low-level loops that would be candidates for vectorization. Particle blocking and sorting have been implemented to increase trip counts of low-level loops and improve memory locality, and OpenMP directives have been added to vectorize compute-intensive loops that were identified by Advisor. The optimizations have improved the performance of the pushe kernel 2x on Haswell processors and 1.7x on KNL. The KNL node-for-node performance has been brought to within 30% of a NERSC Cori phase I Haswell node and we expect to bridge this gap by reducing the memory footprint of compute intensive routines to improve cache reuse. ## PICSAR is a Fortran/Python high-performance Particle-In-Cell library targeting at MIC architectures first designed to be coupled with the PIC code WARP for the simulation of laser-matter interaction and particle accelerators. PICSAR also contains a FORTRAN stand-alone kernel for performance studies and benchmarks. A MPI domain decomposition is used between NUMA domains and a tile decomposition (cache-blocking) handled by OpenMP has been added for shared-memory parallelism and better cache management. The so-called current deposition and field gathering steps that compose the PIC time loop constitute major hotspots that have been rewritten to enable more efficient vectorization. Particle communications between tiles and MPI domain has been merged and parallelized. All considered, these improvements provide speedups of 3.1 for order 1 and 4.6 for order 3 interpolation shape factors on KNL configured in SNC4 quadrant flat mode. Performance is similar between a node of cori phase 1 and KNL at order 1 and better on KNL by a factor 1.6 at order 3 with the considered test case (homogeneous thermal plasma).

97 MATHEMATICS AND COMPUTING↗

Understanding the I/O Performance Gap Between Cori KNL and Haswell

The Cori system at NERSC has two compute partitions with different CPU architectures: a 2,004 node Haswell partition and a 9,688 node KNL partition, which ranked as the 5th most powerful and fastest supercomputer on the November 2016 Top 500 list. The compute partitions share a common storage configuration, and understanding the IO performance gap between them is important, impacting not only to NERSC/LBNL users and other national labs, but also to the relevant hardware vendors and software developers. In this paper, we have analyzed performance of single core and single node IO comprehensively on the Haswell and KNL partitions, and have discovered the major bottlenecks, which include CPU frequencies and memory copy performance. We have also extended our performance tests to multi-node IO and revealed the IO cost difference caused by network latency, buffer size, and communication cost. Overall, we have developed a strong understanding of the IO gap between Haswell and KNL nodes and the lessons learned from this exploration will guide us in designing optimal IO solutions in many-core era.

97 MATHEMATICS AND COMPUTING↗

Basic Energy Sciences Exascale Requirements Review. An Office of Science review sponsored jointly by Advanced Scientific Computing Research and Basic Energy Sciences, November 3-5, 2015, Rockville, Maryland

Computers have revolutionized every aspect of our lives. Yet in science, the most tantalizing applications of computing lie just beyond our reach. The current quest to build an exascale computer with one thousand times the capability of today’s fastest machines (and more than a million times that of a laptop) will take researchers over the next horizon. The field of materials, chemical reactions, and compounds is inherently complex. Imagine millions of new materials with new functionalities waiting to be discovered — while researchers also seek to extend those materials that are known to a dizzying number of new forms. We could translate massive amounts of data from high precision experiments into new understanding through data mining and analysis. We could have at our disposal the ability to predict the properties of these materials, to follow their transformations during reactions on an atom-by-atom basis, and to discover completely new chemical pathways or physical states of matter. Extending these predictions from the nanoscale to the mesoscale, from the ultrafast world of reactions to long-time simulations to predict the lifetime performance of materials, and to the discovery of new materials and processes will have a profound impact on energy technology. In addition, discovery of new materials is vital to move computing beyond Moore’s law. To realize this vision, more than hardware is needed. New algorithms to take advantage of the increase in computing power, new programming paradigms, and new ways of mining massive data sets are needed as well. This report summarizes the opportunities and the requisite computing ecosystem needed to realize the potential before us. In addition to pursuing new and more complete physical models and theoretical frameworks, this review found that the following broadly grouped areas relevant to the U.S. Department of Energy (DOE) Office of Advanced Scientific Computing Research (ASCR) would directly affect the Basic Energy Sciences (BES) mission need. Simulation, visualization, and data analysis are crucial for advances in energy science and technology. Revolutionary mathematical, software, and algorithm developments are required in all areas of BES science to take advantage of exascale computing architectures and to meet data analysis, management, and workflow needs. In partnership with ASCR, BES has an emerging and pressing need to develop new and disruptive capabilities in data science. More capable and larger high-performance computing (HPC) and data ecosystems are required to support priority research in BES. Continued success in BES research requires developing the next-generation workforce through education and training and by providing sustained career opportunities.

97 MATHEMATICS AND COMPUTING↗

Improved treatment of exact exchange in Quantum ESPRESSO

Here, we present an algorithm and implementation for the parallel computation of exact exchange in Quantum ESPRESSO (QE) that exhibits greatly improved strong scaling. QE is an open-source software package for electronic structure calculations using plane wave density functional theory, and supports the use of local, semi-local, and hybrid DFT functionals. Wider application of hybrid functionals is desirable for the improved simulation of electronic band energy alignments and thermodynamic properties, but the computational complexity of evaluating the exact exchange potential limits the practical application of hybrid functionals to large systems and requires efficient implementations. We demonstrate that existing implementations of hybrid DFT that utilize a single data structure for both the local and exact exchange regions of the code are significantly limited in the degree of parallelization achievable. We present a band-pair parallelization approach, in which the calculation of exact exchange is parallelized and evaluated independently from the parallelization of the remainder of the calculation, with the wavefunction data being efficiently transformed on-the-fly into a form that is optimal for each part of the calculation. For a 64 water molecule supercell, our new algorithm reduces the overall time to solution by nearly an order of magnitude.

97 MATHEMATICS AND COMPUTING↗

A structure preserving Lanczos algorithm for computing the optical absorption spectrum

We present a new structure preserving Lanczos algorithm for approximating the optical absorption spectrum in the context of solving full Bethe-Salpeter equation without Tamm-Dancoff approximation. The new algorithm is based on a structure preserving Lanczos procedure, which exploits the special block structure of Bethe-Salpeter Hamiltonian matrices. A recently developed technique of generalized averaged Gauss quadrature is incorporated to accelerate the convergence. We also establish the connection between our structure preserving Lanczos procedure with several existing Lanczos procedures developed in different contexts. Numerical examples are presented to demonstrate the effectiveness of our Lanczos algorithm.

97 MATHEMATICS AND COMPUTING↗

Optimizing Excited-State Electronic-Structure Codes for Intel Knights Landing: A Case Study on the BerkeleyGW Software

We profile and optimize calculations performed with the BerkeleyGW code on the Xeon-Phi architecture. BerkeleyGW depends both on hand-tuned critical kernels as well as on BLAS and FFT libraries. We describe the optimization process and performance improvements achieved. We discuss a layered parallelization strategy to take advantage of vector, thread and node-level parallelism. We discuss locality changes (including the consequence of the lack of L3 cache) and effective use of the on-package high-bandwidth memory. We show preliminary results on Knights-Landing including a roofline study of code performance before and after a number of optimizations. We find that the GW method is particularly well-suited for many-core architectures due to the ability to exploit a large amount of parallelism over plane-wave components, band-pairs, and frequencies.

97 MATHEMATICS AND COMPUTING↗

OpenMP Parallelization and Optimization of Graph-Based Machine Learning Algorithms

In this paper, we investigate the OpenMP parallelization and optimization of two novel data classification algorithms. The new algorithms are based on graph and PDE solution techniques and provide significant accuracy and performance advantages over traditional data classification algorithms in serial mode. The methods leverage the Nystrom extension to calculate eigenvalue/eigenvectors of the graph Laplacian and this is a self-contained module that can be used in conjunction with other graph-Laplacian based methods such as spectral clustering. We use performance tools to collect the hotspots and memory access of the serial codes and use OpenMP as the parallelization language to parallelize the most time-consuming parts. Where possible, we also use library routines. We then optimize the OpenMP implementations and detail the performance on traditional supercomputer nodes (in our case a Cray XC30), and test the optimization steps on emerging testbed systems based on Intel’s Knights Corner and Landing processors. We show both performance improvement and strong scaling behavior. Finally, a large number of optimization techniques and analyses are necessary before the algorithm reaches almost ideal scaling.

97 MATHEMATICS AND COMPUTING↗

Low rank approximation in G 0 W 0 calculations

The single particle energies obtained in a Kohn-Sham density functional theory (DFT) calculation are generally known to be poor approximations to electron excitation energies that are measured in tr ansport, tunneling and spectroscopic experiments such as photo-emission spectroscopy. The correction to these energies can be obtained from the poles of a single particle Green’s function derived from a many-body perturbation theory. From a computational perspective, the accuracy and efficiency of such an approach depends on how a self energy term that properly accounts for dynamic screening of electrons is approximated. The G 0 W 0 approximation is a widely used technique in which the self energy is expressed as the convolution of a noninteracting Green’s function (G 0 ) and a screened Coulomb interaction (W 0 ) in the frequency domain. The computational cost associated with such a convolution is high due to the high complexity of evaluating W 0 at multiple frequencies. In this paper, we discuss how the cost of G 0 W 0 calculation can be reduced by constructing a low rank approximation to the frequency dependent part of W 0 . In particular, we examine the effect of such a low rank approximation on the accuracy of the G 0 W 0 approximation. We also discuss how the numerical convolution of G 0 and W 0 can be evaluated efficiently and accurately by using a contour deformation technique with an appropriate choice of the contour.

97 MATHEMATICS AND COMPUTING↗

Numerical integration for ab initio many-electron self energy calculations within the GW approximation

We present a numerical integration scheme for evaluating the convolution of a Green's function with a screened Coulomb potential on the real axis in the GW approximation of the self energy. Our scheme takes the zero broadening limit in Green's function first, replaces the numerator of the integrand with a piecewise polynomial approximation, and performs principal value integration on subintervals analytically. We give the error bound of our numerical integration scheme and show by numerical examples that it is more reliable and accurate than the standard quadrature rules such as the composite trapezoidal rule. We also discuss the benefit of using different self energy expressions to perform the numerical convolution at different frequencies.

71 CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSIC↗