Developer Guide

Developer Guide for Intel® oneAPI Math Kernel Library Linux*

ID 766690
Date 3/31/2023
Public

A newer version of this document is available. Customers should click here to go to the newest version.

Document Table of Contents

Overview of the Intel® Distribution for LINPACK* Benchmark

The Intel® Distribution for LINPACK* Benchmark is based on modifications and additions to High-Performance LINPACK (HPL) (http://www.netlib.org/benchmark/hpl/) from Innovative Computing Laboratories (ICL) at the University of Tennessee, Knoxville. The Intel® Distribution for LINPACK Benchmark can be used for TOP500 runs (see http://www.top500.org) and for benchmarking your cluster. To use the benchmark you need to be familiar with HPL usage. The Intel® Distribution for LINPACK Benchmark provides some enhancements designed to make the HPL usage more convenient and to use Intel® Message-Passing Interface (MPI) settings to improve performance.

The Intel® Distribution for LINPACK Benchmark measures the amount of time it takes to factor and solve a random dense system of linear equations (Ax=b) in real*8 precision, converts that time into a performance rate, and tests the results for accuracy. The benchmark uses random number generation and full row pivoting to ensure the accuracy of the results.

Intel provides optimized versions of the LINPACK benchmarks to help you obtain high LINPACK benchmark results on your systems based on genuine Intel processors more easily than with the standard HPL benchmark. The prebuilt binaries require Intel® MPI library be installed on the cluster. The run-time version of Intel MPI library is free and can be downloaded from https://www.software.intel.com/content/www/us/en/develop/tools.html .

The Intel package includes software developed at the University of Tennessee, Knoxville, ICL, and neither the University nor ICL endorse or promote this product. Although HPL is redistributable under certain conditions, this particular package is subject to the Intel® oneAPI Math Kernel Library (oneMKL) license.

Intel® oneAPI Math Kernel Library provides prebuilt binaries that are linked against Intel MPI libraries either statically or dynamically. In addition, binaries linked with a customized MPI implementation can be created using the Intel® oneAPI Math Kernel Library MPI wrappers.

NOTE:

Performance of statically and dynamically linked prebuilt binaries may be different. The performance of both depends on the version of Intel MPI you are using. You can build binaries statically or dynamically linked against a particular version of Intel MPI by yourself.

HPL code is homogeneous by nature: it requires that each MPI process runs in an environment with similar CPU and memory constraints. The Intel® Distribution for LINPACK Benchmark supports heterogeneity, meaning that the data distribution can be balanced to the performance requirements of each node, provided that there is enough memory on that node to support additional work. For information on how to configure Intel® oneAPI Math Kernel Library to use the internode heterogeneity, seeHeterogeneous Support in the Intel® Distribution for LINPACK Benchmark.

Product and Performance Information

Performance varies by use, configuration and other factors. Learn more at www.Intel.com/PerformanceIndex.

Notice revision #20201201