The chip manufacturer has announced an OpenGL and OpenCL debugger, profiler and memory analyzer alongside a range of other tools, and competitions with top prizes of $20,000.
gDEbugger was originally developed by Graphic Remedy and has been around since 2004. AMD acquired Graphic Remedy last October, and this latest version moves to run under Visual Studio and under the AMD banner.
AMD has also announced a number of other tools including a parallel path analyzer, an API for Global Memory for Accelerators, and a Task Manager API. All the tools are due for beta release in Q3 this year.
In the meantime, the original gDEbugger can be downloaded and used for free. It lets you debug OpenCL kernels running on AMD GPUS, step into the kernel execution directly from the API call that issues it, and while stepping through the code while examining kernel variables and data. You can view compute and graphic memory objects such as OpenCL images to view their contents as they change from write, copy, and kernel operations. You can also find out which objects are causing memory leaks.
The advantage this offers over separate debuggers for the CPU and GPU is that it provides a single debugging environment for the heterogeneous environment.
The full range of tools is part of AMD’s push to attract the developer community to write for AMD chips rather than Intel. Parallel Path Analyzer (PPA) is a profiling tool that visualizes data transfers and kernel execution, identifies system-wide critical paths and locates data dependencies. The Global Memory for Accelerators API gives you a way to create applications without having to explicitly manage multiple data buffers across the separate address spaces of GPU and CPU. Finally, the Task Manager API lets you manage compute tasks in a heterogeneous multi-core environment with dynamic load balancing.
Alongside the new tools, AMD has announced the AMD OpenCL Coding Competition hosted by Topcoder. The competition aims to add heterogeneous computing to the mix and has a combined total of $50,000 available in prizes.
According to the competition blurb, with AMD Fusion Accelerated Processing Units (APUs) every CPU will have a GPU on the same die, with many functional processors at your disposal without the traditional bandwidth bottleneck of other heterogeneous computing platforms.
The competition has two main components - an innovation challenge and a performance challenge. For the innovation challenge, you need to use OpenCL and the AMD APP SDK to build something that will demonstrate the power of Fusion APUs. The entries can be apps, games, widgets, benchmarks or demos, and there’s a $20,000 first prize. The performance challenge involves a head to head competition with other programmers to optimize an OpenCL algorithm to see who can get the fastest results, again for a $20,000 first prize. Entry details are available from Topcoder.
AMD Developer Central
If you would like to be informed about new articles on I Programmer you can either follow us on Twitter or Facebook or you can subscribe to our weekly newsletter.