OpenCL in Beta SDK for GPUs and Multicore CPUs

Aug 07, 2009

Chipmaker AMD has integrated the open standard OpenCL specification for parallel programming into its ATI Stream development environment.

The Open Computing Language (OpenCL) should allow developers to address both graphical and multicore processors for their parallel programming projects with a new beta version of the ATI Stream SDK version 2.0. The idea behind the OpenCL specification is to harmonize parallel programming for GPUs and multicore CPUs to improve speed and responsiveness.

By AMD's own reckoning, integrating the specification into its ATI development environment represents the first ever OpenCL development platform. The final version of ATI Stream SDK 2.0 is planned for later in 2009. The SDK is downloadable from the AMD project page as binaries in 32-bit and 64-bit versions. The supported Linux distros are openSUSE 11 and Ubuntu 8.04. SDK use requires registration as an AMD developer, after which participation in their forum is available.

OpenCL is an open standard for parallel programming for multicore and graphics processors. The current specification version 1.0 has information on the C-based programming language and the platform-layer and runtime APIs. Around 30 firms are involved in its development, among them IBM, Intel, Toshiba, Freescale, Nvidia, QNX and Texas Instruments. The AMD ATI website includes an introduction to OpenCL. The Khronos Group, which is initiating the OpenCL development effort, has the full 300-page specification on their website.

Related content

comments powered by Disqus
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters

Support Our Work

Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

Learn More