NVIDIA unveils Tesla K40 accelerator, teams with IBM on GPU-based supercomputing

NVIDIA unveils Tesla K40, teams with IBM on supercomputing in the data center

NVIDIA’s Tesla GPUs are already mainstays in supercomputers that need specialized processing power, and they’re becoming even more important now that the company is launching its first Tesla built for large-scale projects. The new K40 accelerator only has 192 more processing cores than its K20x ancestor (2,880, like the GeForce GTX 780 Ti), but it crunches analytics and science numbers up to 40 percent faster. A jump to 12GB of RAM, meanwhile, helps it handle data sets that are twice as big as before. The K40 is already available in servers from NVIDIA’s partners, and the University of Texas at Austin plans to use it in Maverick, a remote visualization supercomputer that should be up and running by January.

As part of the K40 rollout, NVIDIA has also revealed a partnership with IBM that should bring GPU-boosted supercomputing to enterprise-grade data centers. The two plan on bringing Tesla GPU support to IBM’s Power8-based servers, including both apps and development tools. It’s not clear when the deal will bear fruit, but don’t be surprised if it turbocharges a corporate mainframe near you.

Filed under: , ,

Comments

Source: NVIDIA

A Mind-Boggling Display of Earth-Like Planets

Kepler keeps finding more and more potentially habitable planets in our Universe—and it turns out that looking at them can be just as perplexing as thinking about them.

Read more…


    



NASA halts efforts to repair Kepler space telescope

Kepler calls it quits: NASA halts efforts to repair damaged space telescope

It’s had a good run, but it seems like NASA’s Kepler telescope is down for the count — the space agency says it has stopped repair efforts. The 0.95 meter diameter space telescope launched four years ago, tasked with seeking out Earth-sized planets suitable for habitation. All was going well until the rig’s gyroscopic reaction wheels began to fail, robbing it of the precision aim needed to continue its task. After months of testing, NASA has concluded that it won’t be able to restore the telescope to full working order.

That doesn’t mean the mission is at an end, however — NASA still has to sort troves of previously collected data, thumbing through over 3,500 exoplanet candidates to add to the 135 celestial bodies Kepler has already identified. The hardware may one day see a second life too, as engineers attempt to assess what can be done with the remaining two reaction wheels and the telescope’s attitude control thrusters. Without significant (and now abandoned) repair efforts, Kepler will never be precise enough to continue its primary mission, but NASA is hopeful it will eventually find a new purpose.

Filed under:

Comments

Nvidia Brings Kepler To Mobile, Offers Same Graphics Power As iPad 4 With One-Third The Battery Drain

Screen Shot 2013-07-24 at 8.06.41 AM

At the SIGGRAPH conference going on this week, Nvidia has made a potentially huge announcement regarding the future of mobile gaming: the company is bringing its Kepler graphics architecture to mobile devices via its Project Logan next-gen mobile processor. Nvidia compares this development to the rollout of the first GPU, the GeForce 256, 14 years ago. Logan as a platform, with its Kepler GPU, jumps mobile computing ahead by the equivalent of seven years’ worth of advancement, says Nvidia.

Mobile devices haven’t had GPUs available with true, full desktop feature set support before now, Nvidia says. That includes things like better rendering and simulation techniques, like tessellation, advanced physics and anti-aliasing and the ability to calculate lighting effect rendering in a single pass. All of which is pretty technical, but ultimately means that a lot of the tricks and capabilities available to console gaming will now be brought to mobile devices.

Kepler is already in use in desktop GPUs, and in Nvidia’s desktop designs it can take on general purpose computing as well to help with the processing workload even when you’re not playing Call of Duty. Kepler on mobile can also offer this, which means that mobile apps featuring things like computational imaging, computer vision, AR or speech recognition would be able to benefit from the Kepler GPU and take advantage not only of its processing power, but also its power efficiency.

To demonstrate Kepler’s power, Nvidia released a video which showed a realistic human head model being generated in real-time. The demo itself isn’t new – it was shown off earlier this year on a desktop PC using Kepler. But this time around, the mobile Logan processor with Kepler is powering the rendering, which makes this a pretty stark proof of the kind of effect that Kepler could have on the state of mobile gaming, and mobile computing overall.

Many believe that mobile and desktop gaming are not on a collision course – each will have its place, and serve different functions owing to different graphics capabilities, control schemes and more. But with Kepler, Nvidia is showing that it’s delivering the technology to enable a further blurring of the line between desktop and mobile, and delivering it well ahead of when some people thought that might happen.



NVIDIA puts Project Logan on display at SIGGRAPH: Kepler gets cozy on a mobile chip (video)

NVIDIA details Project Logan at SIGGRAPH Kepler gets cozy on a mobile chip video

We’ve known about NVIDIA’s plans to bring Kepler to mobile for a few months now, but the component maker offered up an early glimpse of the SoC at SIGGRAPH this week. In terms of power usage, Logan’s use of Kepler architecture translates to one-third the consumption of GPUs currently running in devices like the Retina iPad while wrangling the same renders. Of course, it does have a healthy amount of room to scale up from there for much beefier tasks. The silicon also supports the just announced OpenGL 4.4, OpenGL ES 3.0 and Microsoft’s DirectX11. So, what does all of that translate to in terms of graphics? Project Logan enables the use of advanced rendering and simulation techniques to construct imagery — things like tessellation, advanced lighting and physical simulation, just to name a few. In other words, this allows for PC and console-quality graphics to get cozy on mobile devices. For a look at chip in action, venture on past the break where the Ira demo that was unveiled earlier this year on GeForce GTX Titan GPU-packing desktop is now running on a Logan-equipped mobile device.

Filed under: , , , ,

Comments

Source: NVIDIA

NVIDIA unveils GeForce GTX 760, brings modern Kepler down to $249 (video)

NVIDIA unveils GeForce GTX 760, brings modern Kepler down to $249 video

NVIDIA has been gradually lowering the base pricing for its desktop GeForce 700 series, but few outside of the hardcore gamer set would say the $399 GTX 770 was affordable. Enter the GeForce GTX 760: the Kepler-based chipset supports all the visual effects of its faster cousins, but at a more palatable $249 target price. Although it won’t rival the 770 in performance, it offers more bang for the buck than the GTX 660 it’s built to replace: the GTX 760 carries more processing cores (1,152 versus 960) and more memory bandwidth (192GB/s versus 144GB/s) while maintaining similar clock speeds. It can even punch above its weight class, as it’s reportedly up to 12 percent faster than the $299 GTX 660 Ti. Should that balance of price and performance sound especially sweet, you can pick up a GTX 760 board today from the likes of ASUS, EVGA, Gigabyte and others. Several PC builders, such as Falcon Northwest, Maingear and Origin PC, are also equipping their machines with the new mid-tier graphics from day one.

Filed under: , , ,

Comments

Source: NVIDIA

NVIDIA to license graphics tech to other companies, starting with Kepler

NVIDIA to license graphics tech to other companies, starting with Kepler

To use NVIDIA’s graphics technology, you’ve typically had to buy gadgets using NVIDIA chips — good for the company’s bottom line, but not for influencing the industry as a whole. The firm is expanding its ambition today with plans to license some of that technology on a broader scale. Beginning with the Kepler architecture, other firms can use NVIDIA’s GPU cores and graphics-related patents for their own processors and chipsets. The deal could affect a wide range of hardware, but it mostly pits NVIDIA against the likes of Imagination Technologies: a system-on-chip designer could integrate a Logan-based GPU instead of the PowerVR series, for example. While it will be some time before third-party silicon ships with NVIDIA inside, it’s already clear that the company’s in-house design is now just one part of a larger strategy.

Filed under: , , ,

Comments

Source: NVIDIA

NVIDIA reveals GeForce GTX 700M series GPUs for notebooks, we go eyes-on

NVIDIA reveals GeForce GTX 700M series GPUs for notebooks -- we go eyes-on

We’ve already seen a couple of new desktop GTX cards from NVIDIA this month, and if the mysterious spec sheet for MSI’s GT70 Dragon Edition 2 laptop wasn’t enough of a hint, the company’s got some notebook variants to let loose, too. The GeForce GTX 700M series, officially announced today, is a quartet of chips built on the Kepler architecture. At the top of the stack is the GTX 780M, which NVIDIA claims is the “world’s fastest notebook GPU,” taking the title from AMD’s Radeon HD 8970M. For fans of the hard numbers, the 780M has 1,536 CUDA cores, an 823MHz base clock speed and memory configs of up to 4GB of 256-bit GDDR5 — in other words, not a world apart from a desktop card. Whereas the 780M’s clear focus is performance, trade-offs for portability and affordability are made as you go down through the 770M, 765M and 760M. Nevertheless, the 760M is said to be 30 percent faster than its predecessor, and the 770M 55 percent faster.

All of the chips feature NVIDIA’s GPU Boost 2.0 and Optimus technologies, and work with the GeForce Experience game auto-settings utility. The 700M series should start showing up in a host of laptops soon, and a bunch of OEMs have already pledged their allegiance. Check out a video with NVIDIA’s Mark Avermann after the break, where he shows off a range of laptops packing 700M GPUs, and helps us answer the most important question of all: can it run Crysis? (Or, in this case, Crysis 3.)

Filed under: , , ,

Comments

Plex releases 3.0 overhaul for Android, 3.2 update for iOS

Plex releases 30 overhaul for Android, iOS 32 update with remote playback

Plex’s Android app revamp has been brewing for awhile, but it’s at last ready: the 3.0 app is out of beta and available for everyone. The remake provides a much more polished interface, PlexSync support and speedier access to large libraries. It’s facing a rocky start, however. The initial 3.0 release required a myPlex account and didn’t include a remote control widget, and those have only just been fixed with a quick follow-up patch. We wouldn’t lean on earlier versions of Android, regardless of what features you like — the interface rewrite cuts off support for OS releases before Android 3.2.

iOS users aren’t left out of the upgrades. Version 3.2 isn’t as dramatic a makeover, but it does offer tangible improvements over 3.1 that include the Android version’s faster media access and fixes for conspicuous PlexSync bugs. Quick updaters even get a reward for their trouble: the 3.2 client lets the iOS app serve as a remote playback target for other Plex-equipped devices. Whichever platform you prefer, the app update (or a fresh $5 copy) is waiting at one of the source links.

Filed under: , , , ,

Comments

Via: Plex (1), (2), (3)

Source: Google Play, App Store

NVIDIA releases GeForce GTX 780 for $649, claims more power with less fan noise

NVIDIA releases GeForce GTX 780 for $649 still Kepler silicon, but more of it

It’s well over a year since the GTX 680 came out, but given how that card was a strong contender it may feel too early for an upgrade. NVIDIA knows the score, which is why it’s made a particular point of pitching this year’s card at owners of the GTX 580 instead. Upgraders from that GPU are pledged a 70 percent lift in performance, which is about double the gain a GTX 680 owner would see. On the other hand, something more people might notice — if NVIDIA’s slides prove to be accurate — is a 5dBA drop in noise pollution, as well a new approach to fan control that attracts less attention by varying revs less wildly in response to load. This is surprising given that most of the extra performance in this card stems from more transistors and greater power consumption, but that’s what we’re told. Feel free to hold out for our round-up of independent reviews or read past the break for further details.

Filed under: , ,

Comments