OpenCL 1.1 Specification Released, Enabling Cross-Platform Parallel Programming

The latest version of the OpenCL standard, OpenCL 1.1, has been ratified and made publicly available by the Khronos Group. This open, royalty-free standard allows for parallel programming across modern processors, providing enhanced performance and functionality. The specification is the result of cooperation between industry-leading companies, including AMD, Apple, ARM, and NVIDIA, among others.

Key Takeaways:

• OpenCL 1.1 provides backwards compatibility with OpenCL 1.0, ensuring that existing code investments are protected.
• The specification includes enhancements for parallel programming, enabling programmers to take advantage of heterogeneous parallel processing.
• The C++ wrapper API for OpenCL is now available, allowing for easier adoption of the standard.
• OpenCL 1.1 conformance tests are immediately available, ensuring that implementations meet the standard.
• The OpenCL 1.1 specification has been released 18 months after OpenCL 1.0, allowing for even more effective use of parallel computing resources.
• Neil Trevett, chair of the OpenCL working group, notes that the commercial opportunity driving OpenCL 1.1 implementation is fueled by the industry's cooperation.
• The OpenCL ecosystem is strengthened by the cooperation of companies like AMD, Apple, ARM, Blizzard Activision, and Intel, among others.

Statistics:

• 18 months: the time between the release of OpenCL 1.0 and OpenCL 1.1.
• 20+ companies: the number of industry-leading companies participating in the OpenCL working group.
• 1.1: the version number of the OpenCL specification.
• 1.0: the previous version of the OpenCL specification.

Sources:

• The Khronos Group press release
• www.khronos.org/opencl/
• Euclid Infotech, 2010, distributed by Contify.com