Allinea Software Blog

Wed July 23, 2014 by Beau Paisley

Allinea DDT and Allinea MAP are excellent tools for finding program flaws and performance issues – they are also very helpful for studying codes and coding techniques. In this article I present a handful of optimization techniques and use Allinea MAP to illustrate their benefits. These tips are focused on scientific programmers who want to adopt best practices and the tips are limited to those that:

  • are beneficial across most current architectures and probably on most future architectures,
  • do not require any deep understanding of a given computer architecture, and
  • do not require a refactoring of your code that reduces readability.

1.   Division is expensive

On most machines the division operator is significantly more expensive (i.e. takes many more clock cycles) than all other operators. When possible, refactor your code to not use...

Mon July 14, 2014 by Beau Paisley

When profiling source code for performance issues it is sometimes desirable to profile a subset of your application. Maybe you already have a high-level view of your application’s performance, or maybe you have been tasked with working on a particular routine. Fortunately, with Allinea MAP, this is a simple process. This article provides an illustration of how to sample a subset of a FORTRAN program.

Mon June 9, 2014 by David Lecomber

We all have our objectives for scalability. It may mean hundreds, thousands, or millions of cores to you.

But spare a thought for the software teams.

We need developers and application analysts in order to scale on that new machine. Do we expect only optimization?

In reality, there is a necessary step before this: debugging.

What should every application developer know about debugging at scale?

Tue May 13, 2014 by Beau Paisley

OpenACC is a directive-based programming model that enables offloading code from C, C++, and Fortran to  accelerators such as GPUs.

Fri April 11, 2014 by The Development Team

Whatever the purpose of a HPC system - from running diverse science tasks to having a single-purpose crunching through the same CFD code 24x7 - fundamentally it is nothing without the applications that use it.

Mon February 24, 2014

In days gone by, Moore’s Law meant a new machine brought a near automatic boost in performance. Today, performance scalability comes through hard and sometimes not so hard won battles with your application source code. I prefer the latter kind of battle.

Mon December 16, 2013

Native mode can be both the smart and the lazy way to use the Intel® Xeon Phi™: applications can run “out of the box” with a simple recompilation preparing for the platform, and then the application is run on the coprocessor, from inside the Linux instance in the coprocessor, and independently from the host processor.