Interest in tablets skyrocketed to an all-time high this past week as Apple announced its iPad tablet. Apple is one of NVIDIA's long-time technology partners, and they use NVIDIA technology in products such as the MacBook, iMac, and Apple TV line.
The iPad reaffirms NVIDIA's vision that 2010 is the year of the tablet. The right technology, medium and killer applications are all finally coming together to make this new generation of tablets massively appealing.
Powering many new tablets coming this year, NVIDIA Tegra is incomparable when it comes to delivering the complete web, HD video playback and sharing, uncompromised multimedia and mobile gaming and awesome battery life.
Tablets powered by Tegra support full Adobe Flash Player 10.1 for stutter-free viewing, provide access to Flash-based sites such as Hulu and Netflix, and run the full versions of popular mobile games such as Farmville. Tegra also provides at least 16 hours of HD video playback and 140 hours of audio playback on tablet devices so the power cord can be left at home.
Ars Technica praised Tegra’s capabilities over other tablets.
"I can easily imagine an Android-based tablet that's designed by HTC, powered by NVIDIA's Tegra 2, does 1080p video (vs. iPad's 720p maximum resolution), and has much better 3D gaming performance, but at the cost of two hours of battery life. And if someone made such a thing, I would rather buy it than an iPad. I'd even pay over $500 for it."
Looking at a slew of upcoming tablets, Mashables chimed in.
"Made by Compal, the 7-inch Android 2.0 tablet runs on the new, high-powered next-generation Tegra 2 processor NVIDIA announced at CES. With this chip, a device can support 1080p video playback, yet retain startlingly good battery life."
Not all netbooks are created equal. PC Magazine named the ASUS 1201N Eee PC "the fastest netbook in the market" thanks to its ION GPU and dual-core Atom CPU. That performance translated to faster video encoding, the ability to play games like World of Warcraft and Spore "without restrictions" and flawless HD video.
"1080p video—tested at various bit rates and formats—played back beautifully on the system and over the HDMI interface (with a Samsung LN55A950 LED Display). And all of my testing was done at full screen."
How do ION netbooks compare with CULV and AMD Neo CPUs? You may be surprised.
"Even netbooks that run on an AMD Neo processor (like the MSI U210) or an Intel Pentium CULV (Toshiba Satellite T115-S1105)—processors considered speedier than a single core Atom—fell short of the ASUS 1201N in terms of performance.”
Intel wants you to think that netbooks are weak, to force consumers to spend more money on bigger, faster, full-featured notebooks. But not everyone needs a bigger, faster, more expensive notebook. If all you want is a small and affordable PC that can handle all your basic tasks plus mainstream games, Flash video, and video conversion, ION netbooks are ideal.
Get Hands-On With Parallel Programming
So you're hearing about all the momentum behind GPGPU and you're itching to try it out and write your first parallel application, but it would be great if you had the advice and guidance of someone that knew this stuff inside out.
The world's first textbook on programming massively parallel processors was released, written by former chief scientist and now NVIDIA Fellow Dr. David Kirk and Dr. Wen-mei Hwu who serves at the University of Illinois at Urbana-Champaign as Chair of Electrical and Computer Engineering in the Coordinated Science Laboratory.
Both David and Wen-mei have been teaching parallel programming ever since NVIDIA released the CUDA architecture. The hands-on experience they have gained from delivering countless lectures and working closely with students provides an insight to up and coming developers that no other book can deliver.
Amazon logged more than 1000 orders for the book in the first few hours of it being on sale and leaders in academia have already endorsed the text.
"David Kirk and Wen-mei Hwu are pioneers in this increasingly important field, and their insights are invaluable and fascinating. This book will be the standard reference for years to come." – Hanspeter Pfister, Harvard University
"For those interested in the GPU path to parallel enlightenment, this new book from David Kirk and Wen-mei Hwu is a godsend." – David Patterson, U.C. Berkeley
"David Kirk and Wen-mei Hwu's new book is an important contribution towards educating our students on the ideas and techniques of programming for massively-parallel processors." – Mike Giles, Oxford University
Interested in teaching yourself CUDA C or OpenCL? Order a copy of Programming Massively Parallel Processors: A Hands-on Approach here.
GPUs Help Virginia Tech Win Best Paper Award
Shortening design cycles and getting products to market faster is critical to every type of company - especially those involved in chip design and NVIDIA should know.
Research done by Sandeep Shukla and his team in the electrical and computer engineering department of Virginia Tech is seeking to tackle this challenge using simulation models created using a language called SystemC and NVIDIA Tesla GPUs.
The highly parallel architecture of GPUs has enabled Sandeep's group to speed up their simulations by 30-100 times – a result so impressive that they just won an award at the 15th Asia and South Pacific Design Automation Conference (ASP-DAC) for their paper titled SCGPSim: A Fast SystemC Simulator on GPUs.
"Speeding up simulation of complex hardware models is extremely important for semiconductor electronics industry to produce newer and newer products in shorter times, thus improving the quality of computing and consumer electronics products faster. If such models can be simulated 10 times faster, then if validating a model took 10 days in the past, now it would take one day. This is why faster simulation performance probably attracted the attention of the ASP-DAC '10 awards committee." Shukla said.
NVIDIA announced its next-generation Tegra, the processor for the mobile web, at CES 2010. Lightning-fast web browsing, full 1080p HD video, complete Adobe Flash Player 10.1 acceleration, 3D graphics acceleration and days of battery life are now possible in the new generation of mobile devices, such as tablets.
AnandTech said,
"Honestly, Tegra 2 is one of the most exciting things I've seen at CES - but it's mostly because of its dual Cortex A9 cores. While I'm excited about improving 3D graphics performance on tablets and smartphones, I believe general purpose performance needs improvement. ARM's Cortex A9 provides that improvement."
Icrontic named Tegra as one of its "Best of CES", saying
"Imagine a device the size of a short paperback that can make short work of 10Mbit 1080p video streams. That is the power of NVIDIA's Tegra 2. Based on the world's first dual core ARM Cortex-A9 processor, the device delivers 10 times the performance of your average smartphone with resolutions that are four to eight fold higher, on a mere 500mW of power. That's about 20 times less power than your average PC."
Trusted Reviews in the UK gave Tegra a ringing endorsement.
The tablet was the talk of CES and NVIDIA is driving the tablet revolution with Tegra. Tablets from ASUS, Notion Ink, MSI, Compal, ICD and Quanta all debuted at CES. jkOnTheRun had a chance to see demos of the Tegra-powered tablets and came away impressed.
"Quite a bit from what was shown - so much so, that the new Tegra platform is timed perfectly with what the company is calling 'the year of the tablet.' Based on the brief demos, we should see portable slates with long battery life and potent performance. One such tablet shown was running Android faster than any smartphone or MID that I've ever seen. And that includes the new Nexus One, which I've had a chance to play with for a short while."
In Ukraine ITC.ua was bowled over.
"Speed of encoding, smoothness and clearness of picture exceed all possible expectations and give the impression of unreality - I have never seen before such a good video picture quality on such a small size device! That's what I call a serious bid for victory!”
Combining Silicon Valley technology with German Engineering, NVIDIA and Audi partnered to bring engineered and tested safety technology to current and future Audi cars. Autoblog tested this technology and commented,
Marrying the Internet with your TV, the Tegra-powered Boxee Box by D-Link also debuted to high buzz and impressive reviews. The Boxee Box won the infamous Last Gadget Standing award, as well as "Best of CES" from Popular Science, PC World and LAPTOP, among others. Tom's Guide pointed out what Tegra brings to the Boxee Box.
A common theme for CES 2010 was 3D, 3D and more 3D. NVIDIA has been a pioneering force in 3D and it showed at CES. 3D Blu-Ray players where extremely hot. Also 3D TVs were on display everywhere. ESPN also announced the first 3D network.
"At CES last week, the atmosphere at Sony's 3D gaming area and at NVIDIA's panoramic 3D Vision Surround gaming demo stations was totally different from that at the many booths with 3D TVs on display. Middle-aged regional distributors slip on the same glasses that they had just used to watch animated characters over-emote in Monsters vs. Aliens on Blu-Ray, and looked at the same LED-backlit displays, but now they were grinning like children, belting out things like 'whoa' and 'that's cool'!"
With NVIDIA 3D Vision available and winning several "Best of the Year" awards, most recently being named a Top 10 Gadget for 2009 by Germany's Stern.de.
"More favorable is NVIDIA's version which enables the computer display to show 3D contents. Special glasses and appropriate software make games three-dimensional. With NVIDIA's 3D Vision kit and a compatible 120Hz LCD display games get a third dimension."
At CES NVIDIA took 3D Vision to the next level with 3D Vision surround.
"The NVIDIA booth was home to a mezmerizing setup that consisted of three projectors displaying a racing simulation across a single wide screen. The system consisted of two, soon to be released GF100 cards in SLI. We were told that one graphics card powered one of the projectors while the other GF100 powered the other two displays."
NVIDIA 3D Vision Surround is the world's first consumer, multi-display 3D solution which allows users to span 3D content across 3 high definition monitors or projectors for a truly breathtaking and immersive gaming experience.
"On the other hand, playing Avatar: The Game on NVIDIA's three-monitor setup is completely astonishing. Its use of 3D is more consistent and universally effective than in the movie."
NVIDIA 3D Vision Surround does for 3D PC gaming just like what IMAX 3D does for movies. By spanning 3D across 3 monitors or projectors, NVIDIA 3D Vision kicks it up a notch or two.
"I have been somewhat dubious about AMD's Eyefinity (although, I admit, I need to spend more time with it), in part because the presence of the display bezels tends to ruin the panoramic effect for me. But with the illusion of depth added to the mix, that problem seems to melt away. My visual system seems to filter out the bezels effortlessly, and one is left with the impression of simply looking through a window at a world of depth beyond. The illusion is more than the sum of its parts: better than panoramic displays or 3D Vision alone."
For home theater types, NVIDIA had new Blu-ray 3D software players, including Arcsoft's TotalMedia Theatre 3 and Cyberlink's PowerDVD Ultra and Blu-ray 3D content, including 3D movie trailers for Disney's Toy Story 3, A Christmas Carol, Alice in Wonderland, and more.
"I was a 3D TV skeptic, until I saw a 3D trailer for Toy Story 3 Wednesday night. For so many who have just upgraded to a 2D HDTV, how can the investment in a new big screen—not to mention all the other gear you'll need—be justified, I wondered. Then I entered CyberLink's suite at the Las Vegas Hilton. The company demonstrated an upcoming Blu-ray 3D capable version of its PowerDVD player software on a Mitsubishi WD-65736 (actually one of the more affordable 3D options, at $1,299 online). The 3D trailer of Toy Story 3 began, playing in standard-definition 2D. But then it gloriously burst forth into 3D as good as you see in the most state-of-the-art cinema. The demo took advantage of NVIDIA's 3D Vision glasses, together with an IR transmitter to get the shutter timing right for each eye's 60-Hz signal – a split 120-Hz signal requiring HDMI 1.4 connectors."
NVIDIA also had the world's first sneak peek of YouTube 3D, running on a 3D technology demonstration version of the Adobe Flash Player software that is viewable with NVIDIA 3D Vision glasses in full color.
It's clear that 2010 is poised to be the year of 3D and NVIDIA is leading the charge to bring 3D to the computing masses. NVIDIA proved at CES that 3DVision is the quickest, lowest cost and most compelling way for the consumer to get to 3D.
Consumers Can’t Get Enough ION
NVIDIA ION single-handedly transformed the netbook market by delivering premium PC performance and features in a small form factor, low-powered PCs. Look what it has done for the HP Mini 311.
"Aside from aesthetics and build quality rivaling that of a more expensive computer, ION is definitely the most impressive part of the Mini 311. If you told me a year ago that in one year I could go to the store and pick up a computer for $399 that could handle HD video playback (let alone flash HD playback), I would have called you a liar. Yet here is the Mini 311 with ION, and it can do those things very well. Not only can it handle local HD video playback very well, but with the recent release of Adobe's flash player 10.1, ION can even play flash HD videos at full framerate, which was once a task that would stutter on even powerful machines."
Let’s be honest, if you get a netbook that meets Intel's artificially crippled definition, you are essentially buying the same functionality of an iPhone, but you get a 10-inch screen and no phone. If you get an ION netbook instead, you get a minature, fully functional computer.
Speeding Scientific Discovery With The "Computational Laboratory"
The first thing that pops into your head when you hear the word "laboratory" is probably a scientist, maybe a mad one, lots of test tubes, bunsen burners and maybe even a machine releasing slightly pointless bolts of electricity.
Last week NVIDIA introduced a slightly different kind of laboratory as an integral component of a new initiative called Tesla Bio Workbench. The program consists of a new community site, with links to a dozen applications designed for biological research, such as AMBER and VMD, that have now been ported to CUDA. By using the processing power of NVIDIA GPUs, the computation times for these every day biological research tasks has, in some cases, gone from years to days!
These "computational laboratories" won't negate the need for real experiments done in a "wet laboratory", but the hope is that by speeding up the early stages of research into finding possible drug candidates for example, NVIDIA can help scientists shave off a significant percentage of that development time and help get new drugs to market quicker.
Has Your Shampoo Got CUDA?
It’s not just the medical and life-science community that can benefit from Tesla Bio Workbench - it turns out that making many household products more effective and environmentally friendly is as much a computer problem as it is a balance of chemicals.
Researchers at Temple University are developing a computer simulation model that provides companies like Procter & Gamble with a fast, cost-effective and accurate tool for improving shampoos and liquid detergents. The research hinges on molecules called "surfactants", the agents that determine the cleaning capacity and texture of shampoos, laundry detergents and many other cleaning products. Axel Kohlmeyer of the Institute for Computational Molecular Science at Temple University had this to say on the benefits they are seeing from GPUs.
"We discovered that by adding just two NVIDIA Tesla C1060 GPUs, each node in our newest cluster can do 16 times more work, and thus multiplies our local compute capacity far beyond what we could previously get through the national supercomputing centers."
Axel and the team at Temple were kind enough to let NVIDIA go on campus and film them for a day to learn more about this interesting field. You can check out the video here.
I participated in a recent conference call with NVIDIA, which covered specifics of their next-genetation GF100 graphics architecture. Most of the information that was discussed with be available at NVIDIA's web site including a white paper with detailed information.
The white paper contains a wealth of information, much of which is highly technical. As I read through the white paper, I noted key featues of the GF100 in the bulletized list below.Other web sites will be publishing thourough previews and will be listed at the end of this post.
Based on the Fermi architecture.
"GF" indicates that the chip is a graphics solution based on the Fermi architecture.
"100" denotes that this is the high end part of the "GF" family of GPUs.
Utilizes six 64-bit GDDR5 memory controllers (384-bit total) to facilitate high bandwidth access to the framebuffer.
Implements hardware accelerated DirectX 11 features including tessellation and DirectCompute.
Supports next-generation effects such as ray-tracing, order-independent transparency, and fluid simulations.
Enhances the geometric realism of characters and objects in games.
Contains twice as many CUDA (Compute Unified Device Architecture) cores (512) over the previous architecture.
The number of Render Output Units (ROP) per ROP partition has doubled over the previous architecture.
Improvement in fillrate enables multiple displays to be driven.
Improved ROP compression techniques lessen the performance hit of 8x multisampling antialiasing.
Supports 32x coverage sampling antialiasing (CSAA) - final pixel color is determined using 8 multisamples and 24 coverage samples.
CSAA was enhanced to support alpha-to-coverage on all samples, enabling higher quality rendering of foliage and transparent textures.
Transparent objects can be rendered without presorting (order independent transparency) enabling developers to create levels with complex glass environments.
On-chip L1 (64KB) and L2 (768KB) caches enable high-bandwidth transfer of primitive attributes between the graphics processor and the tessellation unit and between different graphics processors.
The GigaThread engines reduces context switch time to about 20 microseconds making it possible to execute multiple compute and physics kernels for each frame.
Contains an entirely new distributed geometry processing architecture that is implemented using multiple "PolyMorph" engines.
Parallel geometry processing is possibly the single most important GF100 architectural improvement. The ability to deliver setup rates exceeding one primitive per clock while maintaining correct rendering order is a significant technical achievement never before done in a GPU.
Information concerning price, clock speed, power usage and cooling were not provided.
A list of GF100 articles from other web sites can be found in this forum thread.
That is unfortunate for netbooks without ION GPUs, as Tom's Guide points out.
“The weaker Atom processors paired with Ion managed to perform almost as well as the ultraportable did and in some cases better. The only limitation to these machines during our tests was the CPU itself, specifically the Atom. It almost seems like NVIDIA, with ION, is carrying the weaker CPU forward and helping to make an obviously weak netbook nearly as powerful as an ultraportable."
But it is great news for Ion netbooks like the Asus 1201N, per AnandTech.
"As far as Atom netbooks go, the ASUS Eee PC 1201N is currently my favorite of the bunch... But without better IGPs in current CULV laptops, ION tips the scales more in the 1201N's favor. You get a thin and light netbook that excels at being a portable multimedia station."
3D Vision = Best.Of.2009
Not to be outdone by its ION brother, NVIDIA 3D Vision continued its impressive run on "of the year" awards, by adding Legit Reviews to its trophy case.
"Legit Reviews was a little skeptical of the new technology and gave it a 30-day test drive before we posted our analysis. After 30 days of gaming in 3D we were changed forever and even after watching movies like Avatar in 3D it doesn't fully compare to gaming in 3D. We'll admit we are spoiled by getting to play with new technology like this, but not everything is this good."
Sneaky Rene Haas, the General Manager of the notebook group at NVIDIA, did a teaser for our upcoming Optimus technology on the NVIDIA blog.
"NVIDIA Optimus technology works on notebook platforms with NVIDIA GPUs. It is unique to NVIDIA. It is seamless and transparent to the user. Its purpose is to optimize the mobile experience by letting the user get the performance of discrete graphics from a notebook while still delivering great battery life. Look for more details next month."
After that, speculation ran rampant. Sorry, but we'll just have to wait for more info. The only details NVIDIA is providing at this time are on the NVIDIA nTersect blog. Look for more details next month.
Dell + Quadro = World's Fastest Mobile Workstation
When Dell set out to make the fastest mobile workstation ever, they knew they needed fast graphics...and fast workstation graphics means Quadro FX3800.
Quadro FX 3800 is ideal for demanding design professionals on the go and delivers the ultimate professional graphics for your 17" mobile workstation. But does it really?
"The graphics system consists of the just-released NVIDIA Quadro FX 3800M adapter, with 1GB of dedicated RAM and a 650MHz internal clock -- the fastest currently available model of mobile graphics cards."
It really does. With the M6500, you can have it all and take it with you on the go.
Intel's New (Still Bad) IGP Reviewed
Reviews of the highly anticipated Clarkdale and Arrandale platforms from Intel have been trickling out. While they did manage to get more performance on the Integrated Graphics Processor (IGP) front, the fact is the performance is still WAY below MCP79, not to mention discrete graphics. Intel is starting from a ridiculously low base performance. So "50% better" is still 3x-4x worse than even Atom + MCP79 - which is a generation older technology.
"The new Westmere 32nm execution core runs a bit faster with less power and we hope to see better overclocking as we see BIOS maturation. Now for the love of God, please divorce Westmere cores from that piece of junk Intel calls a "Graphics Media Accelerator!"
"Better Still, battery life didn't seem to take a hit even with the extra performance though high-end, high-res gaming was still a lesson in futility when working without a discrete graphics card. Overall, the chip was a welcome addition to the fold, but we got the feeling that the first wave was priced too high and offered too little of a performance increase on the gaming side to really warrant a wholehearted recommendation."
Intel hasn't made much progress in 3D graphics--a bump up from "really, really bad" only got them up to "really bad." And with GPU Computing on the rise, graphics is not just for gaming anymore. It is clear that graphics performance requirements will outpace the performance that Intel integrated graphics will deliver.
AMD OpenCL + Radeon 4XXX = Kaput
Seems that NVIDIA's OpenCL leadership is showing. While consumers are happily using NVIDIA GPUs with Open CL, some folks are having issues with AMD GPUs.
Performance is so bad, (developer Matt) Taylor adds, that his 2.4GHz Core 2 Quad processor outperforms the Radeon "by a factor of two."
AMD later said that Radeon HD4XXX owners are out of luck for OpenCL.
"This is entirely dependent on how you coded the kernel and what OpenCL features you are using. There are known performance issues for HD4XXX series of cards on OpenCL and there is currently no plan to focus exclusively on improving performance for that family. The HD4XXX series was not designed for OpenCL whereas the HD5XXX series was. There will be performance improvements on this series because of improvements in the HD5XXX series, so it will get better, but it is not our focus."
Too bad Radeon owners, OpenCL applications are not for you. NVIDIA is the pioneering force behind using the GPU for computing. Their C with CUDA extensions was an inspiration for OpenCL and other programming interfaces.
