In this week’s edition, you will notice that Heterogeneous Computing is catching on with many applications (both commercial and consumer) harnessing the power of GPU utilizing NVIDIA CUDA technology, chairman of Stanford University’s computer science department Bill Dally becomes NVIDIA’s Chief Scientist and VP (Research), excerpts of dinner interaction with our CEO with reporters and ofcourse something about the unbeatable performance of our enthusiasts GPUs- GeForce GTX295 and GeForce GTX285. Please enjoy reading!

 

GPU Computing is Catching On

While AMD hints at Stream Computing, NVIDIA has hundreds of commercial and consumer applications already reaping the performance benefits of NVIDIA CUDA technology.  UK publication Develop 3D sees GPU computing as a top 10 trend for 2009, coming in at number 4.

 

“The use of high performance graphics cards for tasks other than 3D graphics is on the up. Graphics Processing Units have become just as powerful as CPUs and highly parallel tasks such as simulation and rendering can be offloaded to the GPU or even multiple GPUs. NVIDIA is leading the charge with its CUDA programming language….”

 

InfoWorld has done the math on why CUDA is so valuable to folks with computing problems that were “embarrassingly parallel” -- one for which little or no effort is required to separate the problem into a number of parallel tasks.  Examples sited include 3-D rendering, face recognition, Monte Carlo simulation, particle physics event reconstruction, biological sequence searching, genetic algorithms, and weather modeling.

 

“For embarrassingly parallel problems, for example digital tomography, an under-$10,000 Tesla personal supercomputer can beat a $5 million Sun CalcUA. CUDA makes the parallel programming tractable.”

 

Hardwarezone has a nice overview of CUDA.

 

“As you can see from the few examples provided here, the performance gain from GPGPU is undeniable. Processes that used to take days now take only hours; and those that used to take minutes can now be accomplished in real time. And we have CUDA to thank for all of this, as it has truly unlocked the potential of GPGPU on NVIDIA graphics cards.”

 

And with graphics cards closing in on 2 Teraflops, moving parallel processing from the CPU to the GPU is going in the right direction.

 

 

Stanford's Bill Dally named Chief Scientist

NVIDIA Corporation has appointed Bill Dally, the chairman of Stanford University’s computer science department, Chief Scientist and Vice President of NVIDIA Research. We also named longtime Chief Scientist David Kirk an “NVIDIA Fellow.”

 

Bill is legendary in the computer industry.  He has made fundamental contributions from parallel computing architectures to interconnects to low power design to super fast IOs.  Bill is expected to contribute at all of those levels and more.  And he will take forward David Kirk’s work of building NVIDIA research into one of the most regarded labs in the world.

 

 

Dinner and a Story

This week NVIDIA CEO Jen-Hsun Huang sat down for dinner with a few reporters and was able to discuss a variety of topics with them.  The Wall Street Journal took away from that discussion that NVIDIA is searching for the soul of the new PC.

 

“What is the soul of the new PC?” Huang asks rhetorically, arguing that NVIDIA is positioned to provide the answer. The company has a product that he thinks proves the point–a set of chips called Ion that can substitute for those that Intel sells along with microprocessors such as Atom, its popular new chip for netbooks. An Atom plus an Ion creates a computer motherboard just a few inches on each side, forming the heart of anything from a new netbook to a module that could be plugged in to give a TV computing power.

 

VentureBeat covers how the ION platform will coexist with Intel:

 

“We think Intel’s Atom is a fabulous processor. It’s small, low-power. Why not give it a companion processor from NVIDIA that brings out its full glory. We hope Intel will see the wisdom of letting great innovative products reach the market.”

 

Laptop Magazine also talked about ION with Jen-Hsun:

 

“In a few more months, people will realize that it will be possible to build the Ion platform around Atom that makes it a really fabulous, premium PC experience. I think that this is the beginning of a new trend, and customers can get the full PC experience without spending much more than $399.”

 

 

Heavenly Review

DriverHeaven has published a 23 page 'super review'. It features a review of a Zotac GeForce GTX 295 and a BFG GeForce GTX 285.  But DriverHeaven stepped it up and covered the whole GPU enchilada – Quad SLI vs. Crossfire X, PhysX, Badaboom, Mirrors Edge, Blu-ray playback.  It is one of the most in-depth reviews I've seen and covers a lot of the “Plus” that NVIDIA GPUs bring to graphics.  This is good work that covers all the aspects of a graphics card purchase, and they do not just base their recommendation on simple bar charts in a few games that all GPU makers tune drivers for.  At the end, both the GTX295 and GTX285 win Driver Heaven's 'Heavenly Gold Award - Best In Class' - the highest achievable.

 

 

Chips on Film

The NVIDIA Tesla C1060 gives technical professionals a dedicated computing resource at their desk-side that is much faster and more energy-efficient than a shared cluster in the data center. ComputerTV has created a detailed video that is well presented.

 

Channel Web is running a 5-minute video highlighting the capabilities of the ION platform.

 

 

Maingear Pwns

Maingear got its start much like NVIDIA, in gaming PCs.  Today Maingear is recognized as a premier high performance systems integrator in North America. When Cnet called OEMs for systems in November, some wanted to go with AMD GPUs.  Maingear went with NVIDIA…good move.

 

“Maingear's Ephex is the fastest PC we've reviewed…Editor’s Choice”

 

Maingear is smart enough to know that your multi-GPU computer is only as good as your drivers.  Just ask 3D News:

 

“The battle is over. Four GPUs from AMD couldn’t do much against three GPUs from NVIDIA combined in 3-way SLI. Three NVIDIA GeForce GTX 280 graphics cards easily outperformed their competitor - two HD 4870X2 combined in CrossFireX.”

 

If these editors will sit down and play some games that are not used as benchmarks, they will discover that on top of better performance, SLI scales on more titles.  It just works. 

 

 

Same Story. Different Chip?

John Dvorak has a story about the Intel Itanium and how it almost killed the computer industry.  The scenario sounds eerily familiar.

“Perhaps the idea behind the chip was sound. Intel had decided that the x86 architecture was stale; it had been cloned by AMD on a separate development track, and it needed to be replaced by something completely different. But this notion was probably initiated as much to screw AMD as it was to move the industry forward.

Utilizing trendy ideas of the era—such as RISC and very long instruction words (VLIW)—Intel was convinced it could do something more modern than the creaky x86 architecture (which first emerged in 1978, for God's sake).

Andy Grove figured that Intel could pull an Apple and do what Macs did when that company transitioned from the 68000 to the PowerPC chip: run legacy apps in emulation. It's been done before, after all, and this chip would be so powerful (they thought) that nobody would even notice. No matter that Apple got lucky with its emulator, and that generally emulation sucks.”

 Fun Exercise: search Itanium, replace with Larrabee. Just sayin’….

 

No one knows what Intel will develop. Today, Intel’s GPU product is a set of PowerPoint slides, but they have said that their GPU is based on x86 CPU processors. This is the old style computing architecture. The future of visual computing is the GPU. By spending millions of dollars developing a GPU, Intel is validating that discrete graphics processors will be critically important in the future of the PC. We believe the GPU is the most essential processor of the 21^st century.

 

 

A Moment in Time

Photosynth is a Microsoft technology that we showed off at nVision.  It creates 3D spaces from anyone's 2D photos.  Microsoft worked with CNN and the Inauguration group to create a 'moment' in time. They asked everyone to take one photo of the moment when Obama takes oath. You emailed the photo to CNN, and they created The Moment.  This is another good example of visual computing.