Virtual Institute — High Productivity Supercomputing

SC15 Tutorial - Insightful Automatic Performance Modeling

Date & location

November 15, 2015 - Austin Conference Center, Austin, TX, USA

Presenters

  • Alexandru Calotoiu, Technische Universität Darmstadt
  • Torsten Hoefler, ETH Zürich
  • Martin Schulz, Lawrence Livermore National Laboratory
  • Felix Wolf, Technische Universität Darmstadt

Abstract

Many applications suffer from latent performance limitations that may cause them to consume too many resources under certain conditions. Examples include the input-dependent growth of the execution time. Solving this problem requires the ability to properly model the performance of a program to understand its optimization potential in different scenarios. In this tutorial, we will present a method to automatically generate such models for individual parts of a program from a small set of measurements. We will further introduce a tool that implements our method and teach how to use it in practice. The learning objective of this tutorial is to familiarize the attendees with the ideas behind our modeling approach and to enable them to repeat experiments at home.

Agenda

1. Welcome, tutorial goals
2. Background in performance modeling
3. Our specific modeling method
  • Method description
  • Performance model normal form
  • Workflow
  • 4.
    Choosing the best models - A discussion
    (break)
    5. Hands-on session
  • Experimental setup
  • Tool demonstration
  • Experiments, including scaling processors, scaling input and performance data provided by attendee
  • 6. Wrap up and discussion

    In preparation for the hands-on exercises, participants are encouraged to download and install the latest CUBE distribution on their notebook computer.

    Registration

    Please use the registration site of SC15.

    Contact

    Alexandru Calotoiu
    Technische Universität Darmstadt
    Email: calotoiu@cs.tu-darmstadt.de