I’m a strong advocate of OpenCL for parallel processing computing. When you combine the compute power of multi-core CPUs with their GPUs, a desktop user now has tera-flops of data crunching capabilities rivaling old Cray computers. The only issue is how and when to access this compute power.
The 3 day summit will cover market insights and specific programming techniques on using OpenCL for heterogeneous computing markets. I’ll be writing more on this as I get more details about sessions and topics.
I’ve posted several videos demonstrating Eyefinity technology for the visualization and CAD industries. But I just stumbled across one that is especially compelling. It is a Dell 17” laptop - officially the Dell Precision M6500 mobile workstation - driving 3 displays using the ATI FirePro M7820 Mobility.
The entire 3:17 minute video is worth watching, but if you fast forward to 2:29, you can see what it means to zoom into and manipulate your CAD rendering across multiple screens.
Tom’s Hardware has an in-depth article that looks at not the question of which single-GPU card is the best, but which cards operate best in pairs and trios. The article is for gaming and consumer cards but nonetheless pretty interesting and telling.
The article tests AMD’s Radeon HD 6950 2 GB cards in a three-way CrossFire configuration vs Nvidia’s GeForce GTX 570 1280 MB in a three-way SLI configuration.
From the article:
“CrossFire came out with a huge overall scaling lead over SLI, and removing the one title that didn’t reflect that average would have made the lead even bigger. Superior scaling allowed two mid-priced Radeon HD 6950s to approximate the performance of two higher-cost GeForce GTX 570s, while three HD 6950s took the performance win over three GTX 570s.”
“AMD started out with the highest efficiency at one card, and superb multi-GPU scaling allowed its efficiency to increase with every card we added. When’s the last time adding parts made your machine more efficient?”
I’ve got to say, I would really like to see these same kinds of test run on the professional FirePro graphics models. What about you?
The CRN Test Center recently reviewed the Dell T1500 Precison desktop PC with ATI FirePro v4800 graphics card.
From the review:
“The machine was impressive even before performance tests were run. Just 35 seconds from power-on, the unit was ready to accept keyboard and mouse input. Geekbench performance was outstanding.” ... “We ran the 64-bit version of Geekbench 2.1.11, under which the Precision delivered a top score of 9,286, faster than any other production desktop PC we’ve tested.
Helping propel this performance was the ATI FirePro, which also earns it AutoCAD certification. With 400 streaming cores, this card has enough processors to run a small city.”
3D Professor has a review of the energy efficient, slim, FirePro V5800 covering OpenGL, DirectX11 and OpenCL benchmarking across CAD applications. Although the card has been out almost a year, this review remains relevant given the recent release of the FirePro V5800 DVI targeting the medical markets.
Cut to the chase: as in previous reviews the FirePro V5800 is deemed a great value (street $350) for a mid-range card to use in CAD and visualization.
If you are a SolidWorks user but could not attend all of the session at SolidWorks World 2011, then you will be pleased to find all of the presentations available online in a nicely sortable, easily accessible (read no login required) way.
You can watch the videos and view the PPT presentations. Categories include:
GDC 2011 is in full swing, and although it is a gaming show, many professional products developers attend to learn visualization techniques. This year AMD is presenting several sessions worth checking out.
Efficient Compute Shader Programming (Wednesday 10:30-11:30 Room 301, South Hall)
Direct3D 11 Compute Shaders can be a powerful feature to improve graphics performance, such as post processing. Compute shaders can also be used to improve the performance of tasks traditionally run on the CPU. However, seemingly minor changes to Compute Shader code can have a huge effect on the performance. This talk will explain how an understanding of the underlying hardware architecture can help you make coding decisions that will make your compute shaders perform better. It will also show how AMDs GPU PerfStudio tool can be used to debug and analyze compute shaders to enable the best performance possible.
Optimizing and Debugging Graphics Applications with AMD’s GPU PerfStudio 2.5 (Wednesday 4:30- 5:30 Room 123, North Hall)
AMD’s GPU PerfStudio 2 is a tool for debugging and optimizing graphics applications that is actively being used by a number of game development studios on their next generation DirectX 11 and OpenGL titles. GPU PerfStudio 2 incorporates a powerful Profiler, Frame Capture, Frame Debugger and integrated Shader Debugger which makes it a versatile and effective debugging tool for the game developer. In this talk AMD will demonstrate the latest version of GPU PerfStudio 2 and will highlight advanced DirectX 11 feature support.
Programming for AMD Fusion APUs (Thursday 9:00-10:00 Room 120, North Hall)
Well take a high level programmers view of AMDs new Fusion APUs which have started to appear in numerous PCs from the start of 2011. Well look at the hardwares DirectX 11 capabilities, how to handle scaling performance for such a diverse platform, what to expect from the CPU side and what opportunities the amazing new levels of integration create.
So how do you move open standards along faster? In a word: education.
The OpenCL University Kit introduced by AMD is an easy tool to enable educators to quickly introduce OpenCL learning into their curriculum.
Included in the University Kit is a 13 lecture series, equipped with instructor and speaker notes, as well as code examples where necessary. An advanced understanding of OpenCL is not needed to understand the course materials; students only require a basic knowledge of C/C++ programming. A C/C++ compiler and an OpenCL implementation (such as the AMD APP SDK) are needed to complete the exercises.
AMD today announced support for Microsoft Windows Server 2008 R2 Service Pack 1 (SP1), featuring Microsoft RemoteFX with ATI FirePro professional graphics.
Microsoft RemoteFX virtualizes 2D and 3D graphics, enabling multiple remote workers to access nearly any type of application or screen content, including 3D applications, on a broad range of connected remote devices including PCs, thin clients and network monitors.
“Taking advantage of host-side rendering, ATI FirePro professional graphics lets applications on the host computer run at full speed. Combining fast host-side rendering with traditional client-side rendering, Windows Server 2008 R2 SP1 gives Microsoft VDI customers great deployment flexibility: In a single solution, you get great graphics performance for device-independent remoting, as well as for network-constrained environments.”
FireUser.com is a community resource for CAD, visualization, 3D, video and engineering professionals to learn about the latest acceleration and display technologies and news with a focus on the AMD FirePro workstation graphics line.
