The main reason for the trip was to participate in the China launch of our VISION Technology from AMD retail merchandising approach. Held in Beijing, it was a great opportunity to extend the idea of VISION ― talking directly to the everyday consumer about daily usage (as opposed to the gadget geeks who want to talk about technology), and in one of the fastest-growing consumer markets in the world.

The China team customized the logo and as you can see they added some Chinese characters and as ever they used the language to help bring deeper meaning to the concept of VISION. Translated (they tell me) the symbols mean “see” and “feel”.

Looks (and sounds) good to me.

Following the launch and some customer meetings, we then headed to Shanghai. This was my first visit to the city and I have to say it’s somewhere that I could live. If Beijing is like Washington, D.C. then Shanghai is more like New York. We had a great dinner downtown in an “ex-pat” area but I was not tempted by the American steakhouse ― there are some things we just do better in Texas, and steak is one of them.

While in Shanghai we visited the Shanghai Supercomputer Center. Not only do they have one of the most powerful supercomputers in the world, the Dawning 4000A based on the AMD Opteron™ processor, they have a very excellent museum ― although their “historical” section includes machines I have sold during my career (a sign of age, alas).

4 of the Top 5

Elsewhere in China they are using AMD GPUs to break new records, helping make four out of the top five most powerful supercomputers in the world ones based on AMD technology.

The Tianhe-1, which is the 5^th most powerful supercomputer in the world, is a very large cluster of 5120 ATI GPUs based on the RV770 architecture (the processor you will find in the higher-end ATI Radeon™ HD 4000 series of cards). Not only is this is the first petaflop GPU cluster, it is the first to break into the Top Ten list, as well as AMD’s first large-scale deployment employing ATI Stream Technology in technical applications.  The system also used Intel-based x86 CPUs – which proves that we design our technology to leverage open and interoperable standards.

For the tech-heads among you, the system should be around 563.1 TFLOPS on Linpack (60% from GPUs) and will have a peak performance of 1.2 PFLOPS (80% from GPUs). The cluster’s primary workload will consist of scientific applications such as oil and gas exploration.

I really think this in an interesting example of how “balanced systems” – by which I mean those which combine CPUs and GPUs – are becoming a mainstream solution. Moreover, additional power comes from adding more GPUs and not CPUs. Something gamers already have learned.

By the way, the name “Tianhe” literally means “galaxy” (actually translates to “river in sky”), which is pretty much where the limit is for this next generation of supercomputers.

 Nigel Dessau is senior vice president and chief marketing officer at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such links sites and no endorsement is implied.

The AMD engineering team has posted our ATI Stream SDK v2.0 for OpenCL 1.0.  Now you can have the ability to tap into the full processing performance available from multi-core CPUs and GPUs at the same time, using an industry standard language.  Write it once, use the OpenCL compiler of choice from the different hardware vendors, and you are now running your app across different platforms.  (Notice we’re not stuck on the single vendor, GPU or CPU-only path favored by others in our industry.) 

I’m not the only one who sees the potential of open standards in this arena.  Many of our technology partners do as well.  Listen to David McAllister, director of open source and standards at Adobe:

“Adobe is a strong believer in industry standards; they allow developers to focus on innovation instead of platform support,” said David McAllister, director Open Source and Standards, Adobe.   “The ATI Stream SDK with OpenCL is a solid implementation of a strong standard.”

And Corel:

“Corel plans to integrate OpenCL into our digital media products and we believe it will offer significant benefits to Corel, our customers and partners,” said Jeremy Liang, senior vice president of Digital Media Development, Corel.  “With this new open standard, AMD will be providing drivers to take advantage of the latest CPU and GPU features which we anticipate will improve the overall performance of our digital media applications running on AMD platforms, while simultaneously reducing our development efforts in this area.”

As an industry insider, it’s obvious that the GPU (working together with the CPU) is the next frontier to help dramatically increase the performance of PC and server platforms, especially for some of the most popular applications.   The YouTube and iPhone generation have made distribution and consumption of high definition media pervasive.   Everyone’s a producer and consumer of moving and still images, distributing their content to their spheres of influence worldwide over the Internet, and over the home network.  With OpenCL, now we have an industry standard to accelerate those applications and make them easier and faster to use.

Don’t take my word for it; you can see HP and AMD’s director of ATI Stream technology, Patti Harrell, talk more about how OpenCL influences this trend here. 

My colleague, Simon Solotko, has an excellent companion blog that gives another perspective on the relationship between, OpenCL, AMD’s traditional support for industry standards, and the new ATI Stream SDK v 2.0. 

So, please help us all.  Take up the challenge and do something great with OpenCL.  Grab the ATI Stream SDK v2.0 and leverage the power of AMD CPUs and GPUs working together to accelerate your vision.

 Nigel Dessau is senior vice president and chief marketing officer at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such links sites and no endorsement is implied.

I first learned to edit video at school, maybe 30 years ago, on reel-to-reel Sony black and white video machines. The process was not as disruptive as, say, editing sound tape with a razor blade (see my years at the BBC about 25 years ago) but one error could ruin many hours of great work with one wrong button press.

Since then we have come a long way, and now video editing is a real-time (nearly), computer-enabled marvel.  What back then only a well-funded company could do with a Quantel Paintbox,  now you and I can do on PCs that, even when you go crazy and over-configure, cost less than $1000.

Ironically, the software can now cost more than that, but that is where the power is.

For many years, off-line editing on PCs was a bust for me. It was expensive and the drivers for the video boards were terrible; I used to joke that the time it took to update the drivers was faster than the time it took me to do an editing job. I did love Adobe Premiere (first really good editing software I used) but I hated the PC experience.

Truth is I went Mac!  

But now, happily, I’m at AMD and have a new AMD Phenom™ II 3Gz Quad-core with Dual ATI Radeon™ HD 4870 cards. That’s one heck of a lot of processing power.  But the issue for me was that, from my Mac days, I know FCP keyboard shortcuts in my sleep: since I went Mac – could I go back?

That question may now be irrelevant thanks to two pieces of software and some nifty enablement courtesy of AMD Fusion.

I will write more on Sony Vegas 9 another day (but I LOVE the new interface and it’s much faster) but today I wanted to talk about Adobe Premiere as part of the latest CS4.

The Next Wave of Video Editing

AMD has just announced a beta plug-in for Adobe Premiere Pro CS4 that dramatically improves the performance of a range of complex video editing tasks. The plug-in is the result of an ongoing collaboration between Adobe and AMD engineers designed to take advantage of ATI Stream technology in a way that allocates processing between available system CPU and GPU resources for maximum application performance.

The Adobe Premiere Pro CS4 plug-in draws on the computational power of both the CPU and GPU to improve the program’s performance. ATI Stream technology allows the powerful GPU to be used for more than just graphics processing, resulting in improved general computing power.  This is the industry-changing power of accelerated computing ― the power of Fusion.

In the case of the Adobe Premiere plug-in, a lot of the processing is still being handled by the multi-core CPU, but what is nice here is that I finally get to unlock ATI Stream compute acceleration capabilities sitting on those ATI Radeon HD 4870 graphics cards. Great graphics cards are not just for gamers.

It’s hard for me to give exact measures of the improvement at home ― I just don’t have the measurement tools, but our labs (who have those tools) tell me with the latest Catalyst driver you can see  up to 8 times greater encoding performance. ¹

More specifically, for those looking to build DVDs (or in need of MPEG2 streams) the AMD Fast MPEG2 encoder performs over 178% faster than Adobe default encoder.²  I still like H.264 for quality and compression and the AMD Fast H.264 encoder performs up to 668% faster than Adobe default encoder.

I also dipped into the latest version of Adobe Photoshop and there are lots of ATI Stream accelerated features there too – more on that another time.

Editing has been almost real-time for a few years, and now I’m waiting for rendering and compression to catch up.  In a pre-Stream world it took 8 hours to render my China Olympics video which ended up being 50 GB ― now I need to compress that to fit onto a DVD and onto a Blu-ray. The beta plug-in for Adobe Premiere Pro CS4 is really starting to take huge chunks out of that problem.  

The future really does seem to be about Fusion.

Now if we could just get Blu-ray players to work faster…

Nigel Dessau is senior vice president and chief marketing officer at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such links sites and no endorsement is implied.

¹The Beta plug-in for Adobe® Premiere® Pro CS4 demonstrates significant increases in video encoding performance.  When installed, the plug-in encoded an H.264, 1440×1080i 29.97 frames-per-second, High Quality file in 47.3s; without the plug-in, Adobe Premiere Pro encoded the same file in 372.5s (Custom pre-set based on 1440×1080i 29.97 frames-per-second High Quality where Video Bitrate = CBR 15 Mbps & Audio Bitrate = 128 kbps), demonstrating an almost 8x faster encode time.  The input file size for each comparison was 367 MB.

System Specifications:  AMD Phenom™ II X4 955 3.2GHz processor; 8GB Corsair Dominator CM3X2G1866C9D memory; Sapphire ATI Radeon™ HD4870 1024MB; Windows Vista Ultimate x64 SP1.  Performance of the Adobe Premiere Plug-In will vary based on system configuration, ATI Radeon product, source file and output settings used. 

²The Beta plug-in for Adobe® Premiere® Pro CS4 demonstrates significant increases in video encoding performance.  When installed, the plug-in encoded an MPEG2, 1440×1080i 29.97 frames-per-second, High Quality file in 38.8s; without the plug-in, Adobe Premiere Pro encoded the same file in 108s (Custom pre-set based on 1440×1080i 29.97 frames-per-second High Quality where Video Bitrate = CBR 15 Mbps & Audio Bitrate = 192 kbps), demonstrating over 178% faster encode time. 

We still believe that.

One of the ways Fusion comes to life is in how a PC is much more than an X86 CPU. The various elements of the PC platform coalesce to improve the user experience and enable the system to get its work done, seamlessly. In that same sense, let us consider how the Fusion brand is being made real every day with ATI Stream technology.

Our ATI Stream technology, which is about sharing workloads between the CPU and the GPU, divides software execution between the various compute engines at the platform level.  Today, more and more applications are highly parallel in their execution, either because the task at hand lends itself to this approach or because software programmers want to take full advantage of the increasingly popular multi-core processors, or perhaps a combination of both. 

This may be a behind the scenes, but its impact is likely very real for you. 

A good example is how video encoding and decoding − or “transcoding” as it is known − has become a mainstream, daily event for the YouTube generation. 

It has fast become central to many consumers’ experience to make digital video content available on any screen size and resolution, high definition or standard, regardless of how the original content was formatted.  What many consumers don’t realize is that this is a highly complex problem that can consume huge amounts of compute resources, often requiring hours rather than minutes or seconds to complete.

As it turns out, transcoding is a highly parallel computing problem, with relatively few instructions being hammered against a huge amount of data, something that graphics processors are particularly good at doing.  To that end, we are working closely with ISVs like Cyberlink and ArcSoft to help them embed a video transcoder in their applications that can leverage ATI Stream technology to greatly accelerate transcoding.  We even offer a free, basic utility that performs video transcoding on many of our latest ATI Radeon™ graphics cards, using the latest ATI Catalyst™ 9.5 driver.  Click here for a blog that includes a “how to” video showing how easy it is to work with video files and the ATI Video Converter.

Beyond transcoding, programmable ATI GPUs are rapidly expanding their role in compute platforms to perform diverse tasks, such as taking advantage of their hundreds of cores for protein folding in life-science research, seismic simulations for geophysical exploration, data encryption and many more applications through ATI Stream.  You can see here how financial data analysis is greatly accelerated with a GPU. 

What is particularly exciting about ATI Stream technology is where it is going.  Programming multi-core architectures is getting more standardized and straight-forward through the OpenCL specification, something AMD is already leveraging.   AMD showed an OpenCL demonstration of Havok Cloth™ game physics at the recent Game Developers Conference (GDC). 

And because OpenCL is exactly that − free and open − the barrier to entry for programming multi-core systems is getting lower.

Later this year, AMD plans to release a new ATI Stream Software Development Kit which will make OpenCL widely available for AMD hardware.  Also planned are modifications at the chip-level designed to leverage the OpenCL specification. 

I will talk more about ATI Stream throughout the year with other specific examples of how AMD’s unique combination of CPU and GPUs are supercharging performance.  In the meantime, you can read more about this here.

And as we like to say, The Future is Fusion.

Nigel Dessau is senior vice president and chief marketing officer at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such links sites and no endorsement is implied.

Openness

Openness was a big theme in our discussions at the show.  When we engage with developers, open industry standards are always top of mind ―the industry wants standards that are open and interoperable – we understand this and want to help enable it.

We applaud our friends at Mozilla and Khronos who are working to create a standard for accelerated 3D graphics on the Web. While delivering web-based games with 3D graphics may be in a primitive state today, this could eventually provide the Web with a new visual dimension not only for online gaming but for applications yet to be created. The possibilities are endless and exciting. 

We announced our own 3D graphics initiative last week with a beta release of AMD GPU PerfStudio 2.0, a platform-agnostic development tool for 3D graphics technology. With the tool, debugging is free to all developers, and ― as you should expect ― is based on open standards.

The Visual Experience

Another big theme at GDC this year was the visual experience ― something I predict is only going to continue to gain in importance, and not just for gamers. As CNET keenly pointed out, game play trumps beauty every time. When you’re at the cutting edge of a technology, it’s easy to focus on form over function.

At the same time, putting tools in the hands of developers to create cinema-quality games has delivered great innovation as well ― just look at Ruby.

I spoke last July about the advent of Eye-Definition computing when we launched our Cinema 2.0 initiative. How have we moved the needle since then?

-          Game realism, scale and reach are becoming increasingly advanced. Until recently the technology to deliver cinema-quality video games was just not there. With our ATI Radeon™ HD 4800 series our goal is to help developers make the most realistic games possible.  And with over 1 TeraFLOPS of computing power in a single card, we’ve come much closer to achieving that goal.

-          As AMD’s Neal Robison and Jules Urbach discussed in their GDC session on game physics and realism, developers rely on the total computing resources available to deliver the best gaming experience possible. We’re working with Havok to use their real-time physical simulations on ATI Stream technology to serve up unparalleled user experience. What’s great about this is that developers can access both CPU and GPU compute resources with OpenCL – bringing Fusion to life!

-          We also announced the availability of the ATI FirePro™ V7750 graphics accelerator for the high-end.  For you graphics professionals, I urge you to give this a shot. Large models, shader-intense apps, the ATI FirePro V7750 has its cake and eats it too.  No reason why you shouldn’t as well.

Happy gaming!

Nigel Dessau is senior vice president and chief marketing officer at AMD. His postings are his own opinions and may not represent AMD’s positions, strategies or opinions. Links to third party sites are provided for convenience and unless explicitly stated, AMD is not responsible for the contents of such links sites and no endorsement is implied.