CPU, GPU, FGPA, or DSP: Heterogeneous Computing Multiplies the Processing Power

istock 912207430

Whether your code runs on industry-standard PCs or is embedded in devices for specific uses, chances are there’s more than one processor that you can utilize. Graphics processors, DSPs and other hardware accelerators often sit idle while CPUs crank away at code better served elsewhere. That was the genesis of the Open Computing Language (OpenCL) nearly a decade ago.

OpenCL provides the framework for writing C and C++ code that runs on “all of the above,” and Intel, as a strong supporter of OpenCL, offers the Intel® SDK for OpenCL™ Applications, a comprehensive development environment for writing and optimizing OpenCL applications on Intel platforms. The SDK is part of a vast portfolio of Intel tools for heterogeneous computing, and it supports the offloading of compute-intensive parallel workloads to Intel graphics technology, thanks to an advanced compiler for OpenCL kernels, a runtime debugger, and a code performance analyzer.

What are the benefits of using the Intel SDK for OpenCL Applications? Developers can customize heterogeneous compute apps to accelerate performance and

  • Deliver high-performance image and video processing pipelines
  • Develop proprietary algorithms that run on multiple types of Intel-powered devices
  • Enhance overall system power efficiency
  • Use a single common front end for coding on a broad range of Intel CPUs, GPUs, or FPGAs
  • Speed innovation for competitive advantage working with software tools and libraries like Intel® System Studio, Intel® Media SDK, and OpenVINO™, and optimize system performance and power Intel® VTune™ Amplifier.

The Intel SDK for OpenCL Applications supports a broad range of processing elements including: Iris® Plus, Iris® Pro, Intel® HD Graphics, Intel® Core™, Intel® Xeon®, Pentium®, and Intel® Celeron® processors and GPUs and OpenCL application versions 2.1, 2.0, and 1.2.

Designed with efficiency in mind, the drivers and runtimes for OpenCL applications are packaged separately from the SDK to minimize the size of the installation package. The SDK package contains the development components, and there are several variations and driver versions, depending on the hardware in your configuration.

The SDK also offers a broad range of development tools and operating environments, including: Microsoft Visual Studio 2012 or later and Eclipse Oxygen (4.7) and neon™ framework (4.6). Operating Systems supported are Windows 10 (64-bit desktop mode), CentOS 7.3, Ubuntu 16.04, and Yocto Project.

Best of all, the Intel SDK for OpenCL Applications and the Intel Graphics Driver and OpenCL Runtimes are available at no cost from Intel.

You can get the latest Intel SDK for OpenCL applications as part of Intel® System Studio, which can help you create your own heterogeneous designs that run on Intel Graphics Technology.