AMD and Mentor Graphics Corp. were named by winners consulting firm VDC Research Group Inc. of the firm's "Embeddy" awards for best in show at the 2011 Embedded Systems Conference. AMD took home the hardware Embeddy for its embedded G-Series application processor unit (APU) product. AMD's Radeon E6760 is the first graphics chip for embedded systems to support Microsoft's DirectX 11, the OpenCL standard and up to six simultaneous displays, according to AMD. Mentor Graphics was awarded a software Embeddy for its Embedded Sourcery System Analyzer technology.
RenderStream announced its AMD Radeon based 21.6 teraflop servers and workstations for OpenCL / OpenGL / Brooks based applications and product development. The workstations offer 1,536 stream processors and eight GPUs per system, which provide access to 12,288 cores and 21.6 TFLOPS of aggregate compute power. As an example from information security, the HD 6970 and HD 6990 based VDACTr8 evaluated over 45 billion solutions per second versus 18 billion for the GTX 580 based systems, depending on the implementation.
Intel announced Intel® OpenCL SDK version 1.1 beta conformant with OpenCL™ 1.1 specification. New version includes an alpha preview of SDK implementation for Linux* operating systems. Intel also introduces the new Intel® OpenCL SDK community where OpenCL developers are encouraged to explore and to share advantages of OpenCL workloads found on Intel® Core™ and Intel® Xeon® processors.
Jon Peddie has done up a review and benchmark of the new AMD 6500 series AIBs.
AMD announced the Radeon E6760 embedded discrete graphics processor. The AMD Radeon E6760 GPU is the first of its kind to offer embedded system designers the combination of OpenCL support along with support for six independent displays. The Radeon E6760 is based on a published Khronos Specification, and is expected to pass the Khronos ConformanceTestingProcess. Current conformance status can be found at http://www.khronos.org/conformance.
AMD has released the April update of the Catalyst Drivers for their graphics cards. Performance improvements include a new OpenCL runtime that will improve performance on APUs and also for PCIe transfers between a CPU and discrete GPU.
AMD announced a new collaboration with MulticoreWare, a leader in software solutions and tool development for multi-core and heterogeneous computing environments, to deliver an advanced set of tools for OpenCL™ optimization. The tools development effort accelerates software developers' ability to create and optimize software that fully exploits the unique processing capability of AMD Fusion Accelerated Processing Units (APUs).
OpenCL™ Common Runtime for Linux® on x86 architecture is an OpenCL layered product that improves the OpenCL programming experience by alleviating the programmer from the burden of managing multiple OpenCL platforms and duplicated resources. It is a dynamic shared library that resides between an OpenCL application and one or more OpenCL implementations, such as those developed by AMD and NVIDIA® .
The Common Runtime supports all of the OpenCL v1.1 APIs in a single OpenCL platform consisting of all devices provided by the underlying implementations. This technology provides an integrated environment that can improve application portability as well as simplifying multi-device programming. The Common Runtime has been tested on the IBM System x® iDataPlex™ dx360 M3 with at least one NVIDIA® Tesla™ M2050 running Red Hat Enterprise Linux 5.5. In addition, OpenCL v1.1 implementations from AMD and NVIDIA were used in conjunction with the OpenCL Common Runtime.
AMD has unveiled its lineup of keynote speakers who will address developers at its upcoming AMD Fusion Developer Summit, a list that includes ARM executive John Davies. AMD on Tuesday said that Davies, vice president of technology in ARM's Media Processing division, will discuss ARM’s history of heterogeneous computing, its market strategy and, most of all, its support for AMD’s OpenCL and other open industry standards.
StreamComputing, an independent trainer for using OpenCL on new processors, is available to give lectures at educational institutes. StreamComputing will visit your institute and give one or two hour lectures to students about how processors of the near future will look like, and work. From GPUs to new CPU-extensions, and from Hybrid CPU-GPUs to mobile processors, and how to program them using OpenCL.