Developer Guide

Developer Guide for Intel® oneAPI Math Kernel Library Linux*

ID 766690
Date 7/13/2023
Public

A newer version of this document is available.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

Document Table of Contents

Improving Performance of Your Cluster

To improve cluster performance, follow these steps, provided all required software is installed on each node:

  1. Reboot all nodes.

  2. Ensure all nodes are in identical conditions and no zombie processes are left running from prior HPL runs. To do this, run single-node Stream and Intel® Distribution for LINPACK Benchmark on every node. Ensure results are within 10% of each other (problem size must be large enough depending on memory size and CPU speed). Investigate nodes with low performance for hardware/software problems.

  3. Check that your cluster interconnects are working. Run a test over the complete cluster using an MPI test for bandwidth and latency, such as one found in the Intel® MPI Benchmarks package.

  4. Run an Intel® Distribution for LINPACK Benchmark on pairs of two or four nodes and ensure results are within 10% of each other. The problem size must be large enough depending on the memory size and CPU speed.

  5. Run a small problem size over the complete cluster to ensure correctness.

  6. Increase the problem size and run the real test load.

  7. In case of problems go back to step 2.

Before making a heterogeneous run, always run its homogeneous equivalent first.

Product and Performance Information

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

Notice revision #20201201