Tuesday, 26 March 2013

AMD News Flash

Einstein@Home – Powered by BOINC
Einstein@Home is a volunteer distributed computing project using idle time on computers provided by more than 300,000 volunteers worldwide for astrophysical data analysis. The Einstein@ home team says “Thanks to OpenCL, one of the project’s data analysis applications now runs up to 20 times as fast as before.” Read more about Einstein@Home.



AMD 2013 Developer Summit Call For Proposals deadline extended
You still have a chance to secure a position at the forefront of the heterogeneous technology movement among your peers, academics, the tech media and the financial community. Based on feedback and interest to expand to scope of the conference, we’re extending the CFP deadline & adding additional topics such as Surround Computing and Heterogeneous System Architecture (HSA). Visit the AMD 2013 Developer Summit Blog Web site for more details.

Aviary Releases C++ AMP Optimized SDK
Aviary brings its easy-to-integrate, cross platform functionality to Windows 8. Now developers can add a powerful and intuitive photo editor to their apps in minutes. The functions are performed on highly parallel GPU cores inside the AMD APU, which allows for instant processing. Check out the SDK and come to PhotoHackDay4 on April 6-7th at the Facebook HQ to hack with fellow developers.



AMD CodeXL 1.1 Released on developer.amd.com
AMD CodeXL is a unified developer tool suite that enables you to quickly and easily identify performance issues and programming errors in applications without requiring source code modifications. Version 1.1 improvements include support for Windows 8, Windows Vista, Linux & Kabini APU as well as Visual Studio 2012 integration. Please send your CodeXL feedback to Marty Johnson and team via the CodeXL forum.


Come say "Howdy" during the Game Developers Conference (GDC) 2013
If you’ll be at the GDC 2013 from March 25-29 in San Francisco, you should head over to the AMD booth (#814). You can check out the AMD Eyefinity multi-monitor displays including 3x1 and 5x1 screen configurations. We’ll also be showing off our suite of development tools for optimizing the gaming experience for your customers as well as showcasing some of the best digital content creation tools on AMD FirePro™ Workstation graphics cards.

Check out OpenCV 2.4.4 with OpenCL-based Hardware Acceleration
OpenCV (Open Source Computer Vision) is a library of programming functions for the real-time computer vision. Version 2.4.4 improvements include:


  • interop with OpenCL enables very easy integration of OpenCV in existing OpenCL applications;
  • the ocl::Mat can now use “special” memory (e.g. pinned memory); &
  • the ocl module can detect if the underlying hardware supports “integrated memory,” and if so use “device-local” memory by default for all operations.

From the Twitterverse…
@postgoodism writes “I know PC GPU dev tools are meant to suck and all, but in ~10 minutes w/AMD's CodeXL I fixed a nasty bank conflict & sped up my kernel 5x.” Learn more about CodeXL.




 Training  Date  Location
OpenCL Programming Course by Acceleware April 9-12, 2013 Los Angeles, CA
OpenCL Programming Course by Acceleware October 22-25 2013 Houston, TX




OpenCL Software Development Downloads
OpenCL Training Online
Recommended Reading
University hcResources
Regional hcResources
 


How much do you know about small temporary arrays in OpenCL? drallen writes on the AMD Developer Forums “But beware the devil.  When the array indices are not known at compile time, both gcn and VLIW will access registers serially one thread at a time. gcn scans the 'lanes' for threads looking for an index. It will then read/write all threads with the same index in parallel using v_movereld/s, then repeat until all threads are processed.  Worst case is all 64 indexes in a wave are different = 64 read/write loops (yes branching too). Best case, all indexes are the same and there is only one read/write. (actually that's pretty cool.)  Although VLIW uses A0 register, it also does something similar to serially access different indices.” Learn more on the AMD Developer Forums.

Submit your Tips&Tricks to hcNewsFlash@amd.com




AMD is Hiring!

Featured Job:
Submit Heterogeneous Compute job listings to hcNewsFlash@amd.com





Connect with the heterogeneous community and be a part of this new era of computing!
 

No comments:

Post a Comment