Dr. Dobb's Parallel Articles

The Quiet Revolution in Programming

Wed, 03 Apr 2013 07:14:00 -0400

During the last two years, one of the longest eras in programming has quietly drawn to a close.

Numerical and Computational Optimization on the Intel Phi

Tue, 19 Mar 2013 03:01:00 -0400

How tuning functions for large data sets and profiling the results gets most of the benefits of the Phi's 60 cores without hand wringing and late-night hacking.

Getting to 1 Teraflop on the Intel Phi Coprocessor

Tue, 12 Mar 2013 06:53:00 -0400

The key to truly high performance with the Phi coprocessor is to express sufficient parallelism and vector capability to fully utilize the device. Here is a timing framework that enables you to measure and optimize performance and push it past 1 teraflop.

Concurrent Programming with Chain Locking

Tue, 26 Feb 2013 06:10:00 -0500

Concurrent access to trees and lists requires carefully managed fine-grained locking. Here's a generic solution in C# that removes many of the typical problems.

Task-Based Programming in Windows

Tue, 19 Feb 2013 02:30:00 -0500

Convert programs into a series of independent executable parallel tasks.

Programming the Xeon Phi

Sat, 02 Feb 2013 06:06:00 -0500

A series of articles on getting the best performance out of the new Intel Xeon Phi coprocessor

Comparing OpenCL, CUDA, and OpenACC [video]

Tue, 01 Jan 2013 00:36:00 -0500

Rob Farber takes you on a tour of the paths to massively parallel x86, MultiGPU, and CPU+GPU applications.

Introduction to OpenACC [video]

Mon, 24 Dec 2012 02:27:00 -0500

Using OpenACC to write your first hybrid application

The Best of 2012

Fri, 21 Dec 2012 02:47:00 -0500

The most popular articles of the past 12 months from Dr. Dobb's, plus some additional pieces chosen for your thoughtful consideration by our staff.

CUDA vs. Phi: Phi Programming for CUDA Developers

Mon, 17 Dec 2012 09:17:00 -0500

Both CUDA and Phi coprocessors provide high degrees of parallelism that can deliver excellent application performance. For the most part, CUDA programmers with existing application code have already written their software so it can run well on Phi coprocessors. However, additional work may be required to achieve the highest possible performance.

Introduction to OpenCL [video]

Mon, 17 Dec 2012 00:46:00 -0500

Ben Gaster from AMD Research talks about the design and use of the language OpenCL, which has been embraced by Apple, Intel, and Nvidia among other companies, to accelerate programs.

Programming Intel's Xeon Phi: A Jumpstart Introduction

Mon, 10 Dec 2012 17:07:00 -0500

Reaching one teraflop on Intel's new 60-core coprocessor requires a little know-how

Heterogeneous Programming

Sat, 08 Dec 2012 07:00:00 -0500

AMD's Ben Sander shares details about the heterogeneous system architecture (HSA) and how it will change the way people program in the future.

Introduction to CUDA C/C++

Thu, 29 Nov 2012 07:56:00 -0500

Nvidia's Mark Ebersole introduces core concepts of heterogeneous computing concepts with CUDA C/C++ in this 30 minute tutorial.

Scaling Up And Out

Mon, 19 Nov 2012 02:38:00 -0500

Most attention today is focused on adding nodes or cloud instances to scale out systems. Guest editor Nikita Shamgunov emphasizes the importance of scaling systems vertically as well.

Intel's 50-Core Xeon Phi: The New Era of Inexpensive Supercomputing

Tue, 13 Nov 2012 11:01:00 -0500

The advent of Intel's massively parallel coprocessor will make every server a supercomputer.

Cache-Friendly Code: Solving Manycore's Need for Faster Data Access

Mon, 12 Nov 2012 03:40:00 -0500

As the number of cores in multicore chips grows — Intel just announced the 50+ core Xeon Phi — ensuring that program data can be delivered fast enough to be consumed by so many processors is a huge challenge. Optimal use of processor caches is a key solution, and knowing these coding techniques will become a requirement.

AMD's Bold ARM Server Gambit

Tue, 30 Oct 2012 11:25:00 -0400

By combining 64-bit ARM processors with server-side technology, the company that led the x86 architecture into the 64-bit world is hoping to reinvent the data center and give itself new life.

Creating and Using Libraries with OpenACC

Mon, 29 Oct 2012 11:29:00 -0400

How to write reusable methods (libraries, subroutines, and functions) that can transparently call optimized CPU and GPU libraries using OpenACC pragmas.

Parallel In-Place Merge

Tue, 09 Oct 2012 15:43:00 -0400

Merging sorted arrays in parallel and in place can be done very efficiently, using this algorithm. Comparisons with the performance of similar STL functions are included.

What's New in .NET Framework 4.5

Tue, 02 Oct 2012 07:17:00 -0400

From arrays that can now exceed 2 GB to enhanced background garbage collection, changes in this release of .NET provide immediately useful capabilities.

The OpenACC Execution Model

Mon, 27 Aug 2012 17:32:00 -0400

In this second part of the introduction to OpenACC — the OpenMP-style library for GPU programming — the execution model is explained and samples are benchmarked against straight, OpenMP parallelism.

Parallel Evolution, Not Revolution

Mon, 13 Aug 2012 16:14:00 -0400

Not all parallel programming is fine-grained. But it's still parallel.

Improving Futures and Callbacks in C++ To Avoid Synching by Waiting

Thu, 26 Jul 2012 04:29:00 -0400

In C++, futures are a great way of decomposing a program into concurrent parts, but a poor way of composing those parts into a responsive and scalable program. Microsoft's Parallel Pattern Library (PPL) provides a solution using tasks.

Will Parallel Code Ever Be Embraced?

Wed, 18 Jul 2012 07:16:00 -0400

The advent of the many-core era is not going to push developers to write more parallel code. That hasn't happened as we've gone from 1- to 2- to 4- to 8-core processors, has it? Writing parallel code is not the problem. So, let's stop trying to fix it.