Numerical Scientific Applications Research Articles

Investigating power capping toward energy‐efficient scientific applications

SummaryThe emergence of power efficiency as a primary constraint in processor and system design poses new challenges concerning power and energy awareness for numerical libraries and scientific applications. Power consumption also plays a major role in the design of data centers, which may house petascale or exascale‐level computing systems. At these extreme scales, understanding and improving the energy efficiency of numerical libraries and their related applications becomes a crucial part of the successful implementation and operation of the computing system. In this paper, we study and investigate the practice of controlling a compute system's power usage, and we explore how different power caps affect the performance of numerical algorithms with different computational intensities. Further, we determine the impact, in terms of performance and energy usage, that these caps have on a system running scientific applications. This analysis will enable us to characterize the types of algorithms that benefit most from these power management schemes. Our experiments are performed using a set of representative kernels and several popular scientific benchmarks. We quantify a number of power and performance measurements and draw observations and conclusions that can be viewed as a roadmap to achieving energy efficiency in the design and execution of scientific algorithms.

NumericalPython: A Practical Techniques Approach for Industry

This is a very useful book, especially for Journal of Statistical Software (JSS) readers who wish to learn more about numerical methods and/or scientific applications of Python. Though arguably the book focuses more on numerical analysis than on statistics, it has three solid chapters on the latter, plus abundant infrastructure material useful for statisticians. The author, a physicist by background, has more recently been working in the data science area.

Auto‐tuning of level 1 and level 2 BLAS for GPUs

SUMMARYThe use of high‐performance libraries for dense linear algebra operations is of great importance in many numerical scientific applications. The most common operations form the backbone of the Basic Linear Algebra Subroutines (BLAS) library. In this paper, we consider the performance and auto‐tuning of level 1 and level 2 BLAS routines on graphical processing units. As examples, we develop single‐precision Compute Unified Device Architecture kernels for three of the most popular operations, the Euclidian norm (SNRM2), the matrix–vector multiplication (SGEMV), and the triangular solution (STRSV). The target hardware is the most recent Nvidia (Santa Clara, CA, USA) Tesla 20‐series (Fermi architecture), which is designed from the ground up for high‐performance computing. We show that it is essentially a matter of fully utilizing the fine‐grained parallelism of the many‐core graphical processing unit to achieve high performance for level 1 and level 2 BLAS operations. We show that auto‐tuning can be successfully employed to kernels for these operations so that they perform well for all input sizes. Copyright © 2012 John Wiley & Sons, Ltd.

An object-oriented interface for the dynamic memory management of sparse discrete mathematical operators in numerical scientific applications

This paper describes an object-oriented interface for the memory management of sparse discrete mathematical operators in numerical scientific applications. The basic data structure we propose is intended to implement and cope with mathematical vectors in sparse format. This structure allows an effective implementation of more complex data structures, such as matrix and matrix-like objects, to represent the discrete version of differential operators in numerical application codes. A research example concerning adaptive wavelet algorithms illustrates the possible applications.

Compact DAG Representation and Its Dynamic Scheduling

Scheduling large task graphs is an important issue in parallel computing. In this paper we tackle the two following problems: (1) how to schedule a task graph, when it is too large to fit into memory? (2) How to build a generic program such that parameter values of a task graph can be given at run-time? Our answers feature the parameterized task graph (PTG), which is a symbolic representation of the task graph. We propose a dynamic scheduling algorithm which takes a PTG as an entry and allows us to generate a generic program. We present a theoretical study which shows that our algorithm finds good schedules for coarse-grain task graphs, has a low memory cost, and a low computational complexity. When the average number of operations of each task is large enough, we prove that the scheduling overhead is negligible with respect to the makespan. We also provide experimental results that demonstrate the feasibility of our approach using several compute-intensive kernels found in numerical scientific applications.

Data Processing: Automation in Calculation

A WIDE variety of computational aids is currently available to educators, ranging from slide rule and desk calculator to punched-card equipment and electronic computation. In this review particular attention will be given to electronic computation, for several reasons: (a) The electronic computer will probably be the preferred aid in extensive calculations whenever available; (b) the machines require the more radical changes in research methods; and (c) the area is less known. But since much largescale educational calculating is currently done with punched-card equipment, developments in that field will thereafter be briefly summarized. Much of the computational literature deals with topics, such as numerical analysis, computer logic, computer engineering, scientific applications, and business applications, which possess no specific reference to education or psychology. Instead of trying to make a comprehensive survey of these fields, the policy will be followed of supplying certain key references, generally themselves well documented and of recent date, which provide starting points for any reader wishing to explore further. Sections of the review are bibliographies, journals and professional societies, types of computers, digital computers, computer availability, programing an electronic computer, numerical analysis, computer use in education and psychology, factor analysis, punched-card procedures, other computational aids, and the brain-machine analogy.