AMD's Accelerated Processing Units (APU) present users with a multi-purpose processor that provides superb performance for today's tasks and the potential for even greater performance for tomorrow's applications.

 

An AMD APU incorporates a multi-core CPU and a class-leading AMD Radeon™ HD GPU on the same piece of silicon. The combination of processing units is not only impressive from an engineering point of view but has numerous advantages for the user.

 

For decades AMD's x86 CPUs have been powering general purpose compute and the APU continues that tradition. However as compute is getting more visual and immersive, it requires a high performance GPU alongside a traditional CPU.

 

In the past, GPUs have been used almost solely to push pixels onto a screen. AMD's Radeon family of GPUs have been some of the highest performing GPUs on the market for well over a decade and I believe will continue to power the best visual experiences across multiple form factors in the future.

 

Today however, GPUs can be used for much more than pushing pixels. AMD's A-Series APUs have full-profile OpenCL™ support, an open programming language that allows developers to harness processing power within a GPU to accelerate applications, supplementing the processing power of the APU's CPU cores.

 

AMD has been working with some of the industry's most innovative software developers to foster the use of OpenCL and give both APU users and those with supporting AMD Radeon™ HD and AMD FirePro™ graphics cards the ability to unlock even more performance . One example is the work AMD is doing with Adobe to enable OpenCL acceleration[i] across Adobe® Premiere® Pro CC, Adobe SpeedGrade® CC, and Adobe Media Encoder CC.

 

AMD is a member of The Document Foundation[ii] , the organization that oversees development of the popular LibreOffice suite. AMD is working on incorporating OpenCL support within the LibreOffice Calc spreadsheet application to bring considerable speed-up to complex analysis functions, allowing users to quickly understand vast swathes of data.

 

Adobe and The Document Foundation are just two examples, among many, where AMD is working with software developers on large projects to harness the power of AMD GPUs. However AMD is also giving independent developers the tools needed to easily make use of OpenCL.

AMD recently released ClMath[iii] , an open source OpenCL math library that includes basic linear algebra subprograms (BLAS) and fast Furrier transform (FFT) functions. BLAS and FFT functions are known to be well suited to GPU acceleration and can be used for tasks such as signal processing and image processing, with the ClMath library offering developers the chance to speed up their code without having to invest time to write complex functions.

 

Next month, the AMD Developer Summit[iv] will offer developers a chance to learn about the latest technologies and techniques in order to use the GPU within all AMD APUs and graphics cards to deliver the next level of user experiences. Such is the industry support for AMD's vision in GPU-based compute, keynotes will be given by representatives from some of the industry’s biggest names including ARM, DICE, Imagination Technologies, MediaTek, Oracle, and Unity Technologies along with AMD. The AMD Developer Summit will also feature dozens of talks with technical tracks, allowing conference attendees to see and learn how the industry is embracing GPU compute.

 

It is not surprising that so many big companies are putting their weight behind AMD's leadership of GPU-based computation. GPU compute offers users the chance to experience super-computer levels of processing power from a single AMD APU. AMD is working to enable developers create applications that give users this step change in compute power and with upcoming APUs supporting Heterogeneous System Architecture (HSA)[v] , the GPU is set to become a first-class citizen when it comes to providing all-round compute.

Be at the epicenter of heterogeneous computing with the 2013 AMD Developer Summit. Register for APU13 here.

 

Lawrence Latif is the Manager of Technical Communication at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites, and references to third party trademarks, are provided for convenience and illustrative purposes only.  Unless explicitly stated, AMD is not responsible for the contents of such links, and no third party endorsement of AMD or any of its products is implied.


 


[i] AMD-Adobe Collaboration Unlocks Outstanding Creative Workflows, Enhances Quality and Speed for Video Pros Everywhere - http://www.amd.com/us/press-releases/Pages/amd-adobe-collab-2013sept12.aspx

[ii] AMD joins The Document Foundation Advisory Board to accelerate LibreOffice - The Document Foundation Blog - http://blog.documentfoundation.org/2013/07/03/amd-joins-the-document-foundation-advisory-board-to-accelerate-libreoffice/

[iii] AMD releases APPML source code, creates clMath library - http://developer.amd.com/community/blog/amd-releases-appml-source-code-creates-clmath-library/

[iv] AMD Developer Summit - http://developer.amd.com/apu/

[v] What is Heterogeneous System Architecture (HSA)? - http://developer.amd.com/resources/heterogeneous-computing/what-is-heterogeneous-system-architecture-hsa/