Under What Conditions Will My Application Give Reproducible Results?

Users are often surprised when the same application run on the same data doesn’t give the same result. (WTH?) Welcome to floating-point arithmetic—an inexact and inherently uncertain approximation that’s great … except for when it isn’t. Like for critical functions that require exact reproducibility such as finance, quality assurance, and legal stuff.

This webinar addresses the dilemma and shows you how to build a serial application that delivers repeatable results …

  • When simply rerunning the same binary executable
  • For debug builds and release-optimized builds
  • When targeting or running on different processor types

In addition, learn the conditions under which reproducible results can be obtained for parallel applications making use of the Intel® Compiler with OpenMP* and key Intel® Performance Libraries: Intel® Math Kernel Library, Threading Building Blocks, and Intel® MPI.

Martyn Corden, Technical Consultant Engineer, Intel Corporation

Martyn Corden is a Technical Consulting Engineer in the Developer Products Division at Intel Corporation. He provides technical support for Intel® Fortran and C/C++ compilers for Windows*, Linux* and macOS*, with particular focus on HPC applications. Martyn came to Intel from the Supercomputer Computations Research Institute at Florida State University, where he had extensive experience with high-performance scientific applications. He has many years’ experience in writing, debugging, maintaining, porting and optimizing software for high-energy physics, including for several experiments at CERN in Geneva. Martyn holds a BA in Physics from Oxford University and a Ph.D. in High Energy Physics from the University of Birmingham.

For more complete information about compiler optimizations, see our Optimization Notice.