Most Recent Build Posts RSS



MPI Primer: What You Need to Know

Though microprocessors have become multi-core relatively recently, this happened a lot earlier in the world of mainframes, minis and workstations. If you wanted to write parallel programs on the different architectures used in the science and research fields, it was a bit messy. Back 25 years ago, there was no …

Read Full Post Posted in Build | Tagged , | Leave a comment

Parallel Programming Best Practices Revealed

Intel engineers James Reinders and Jim Jeffers have written a new how-to book that teaches developers to optimize parallel performance on Intel’s multicore and many-core processors. The book, titled “High Performance Parallelism Pearls: Multicore and Many-core Programming Approaches,” provides real-world examples and source code on how to leverage parallelism on …

Read Full Post Posted in Build | Tagged , , | Leave a comment

Intel Software Dev Products Win Top HPC Honors

HPCwire magazine has recognized two Intel clustering technologies as the best in the industry. Intel MPI Library 5.0, which focuses on making applications perform better on Intel architecture-based clusters, won a readers’ choice award for best HPC cluster solution or technology. Intel Parallel Studio XE Cluster Edition — High Performance …

Read Full Post Posted in Build | Tagged , , | Leave a comment

Efficiently Tune Your OpenMP Code to Bust Bottlenecks

Sometimes OpenMP applications don’t run as fast or as efficiently as planned. A newly updated tool by Intel, called the VTune Amplifier XE 2015, can help developers find and fix performance bottlenecks and efficiency problems. The VTune Amplifier XE 2015, part of Intel Parallel Studio XE 2015 Professional Edition, shows …

Read Full Post Posted in Build | Tagged , | Leave a comment

Smooth Out Speed Bumps in C++, Fortran with VTune

You’ve written your high-performance program, but it’s running slower than a three-legged donkey in the Kentucky Derby. Clearly, you need to identify where the problems are, which parts can benefit from being rewritten, etc. Intel’s VTune Amplifier XE is an application for both Windows and Linux that can help you …

Read Full Post Posted in Build | Tagged , , | Leave a comment

Speed Up Image Processing with OpenMP

Image processing can be significantly optimized with the help of OpenMP. Rick Leinecker shows you how, walks you through the OpenMP techniques employed that make the processing faster, and analyzes an iterative OpenMP construct. Since OpenMP applies concurrent processor operations, synchronization is essential to avoid race conditions. In the demonstration …

Read Full Post Posted in Build | Tagged , | Leave a comment

How to Make C++ Run Faster with Vectorization and Parallelization

If you want your C++ program to run significantly faster, turn to vectorization and parallelization. David Bolton walks you through these two high-performance techniques and shows you the steps toward speeding up your code.

Read Full Post Posted in Build | Tagged , , | Leave a comment

High-Performance Fortran: What You Need to Know

High-Performance Fortran (HPF) is an extension to Fortran 90 and later additions in Fortran 95 that you can use to speed up your programs. It was an early parallel technology, developed between 1991 and 1996 and although it is not used so much now, it’s had a lasting effect and …

Read Full Post Posted in Build | Tagged | Leave a comment

How to Avoid Typical OpenMP Traps

In this article I’ve put together a few common OpenMP questions and their answers. If you are new to OpenMP or have a baffling error, I recommend that you read 32 OpenMP Traps For C++ Developers. It’s a great way to classify the most typical traps and help you avoid …

Read Full Post Posted in Build | Tagged | Leave a comment

Performance, Power and Price: Get the Most Out of Xeon Phi

What is the relationship between price-performance and performance vs. power consumption? This Colfax International white paper written by Andrey Vladimirov takes a deep dive into parallel scalability, power and cost factors both with and without Xeon Phi coprocessors. The results are mapped for clusters vs. single devices, with a performance …

Read Full Post Posted in Build | Tagged | Leave a comment