Hasta La Gadget!

Nobody will be too shocked to hear NVIDIA expects its next flagship GPU to be the fastest that’s ever been, but few will have guessed it’d also be one of the company’s coolest and quietest. In the first public teasing of its next-gen graphics card (which is almost certain to bear the GeForce GTX 580 name tag), NVIDIA has revealed a new vapor chamber cooling system, which reminds us of Shuttle‘s ICE CPU cooler — basically, water sealed within the chamber gets boiled by the hot elements (a copper plate in NVIDIA’s case), which forces it to transfer heat away to the bits that are being cooled by the fan, where it chills out and recycles itself back to the boiling plate. The end result, according to NVIDIA, is about seven decibels less vroom relative to the GTX 480, along with lower operational temperatures. Besides that, the company’s Tom Petersen also showed off an impressive tessellation demo and the first public display of Call of Duty: Black Ops gameplay, which was powered by this as yet unannounced GPU. Skip past the break to see it all on video.

[Thanks, Vygantas]

Continue reading NVIDIA promises ‘fastest DX11 GPU on the planet’ very, very soon (video)

NVIDIA promises ‘fastest DX11 GPU on the planet’ very, very soon (video) originally appeared on Engadget on Sun, 07 Nov 2010 09:34:00 EDT. Please see our terms for use of feeds.

Permalink RumorPedia.net  |  nvidia (YouTube)  | Email this | Comments

Add the ‘made in China’ tag to yet another gadget: the world’s fastest supercomputer. China says it has the most powerful computing system — a machine called Tianhe-1A.

The supercomputer uses 7,168 NVIDIA Tesla M2050 GPUs (graphics processing units) and 14,336 Intel Xeon CPUs and is capable of clocking 2.507 petaflops or 2,507 trillion floating point calculations per second.

The Tianhe-1A will take the top spot from the U.S. Cray XT5 aka ‘Jaguar’ that’s at the Oak Ridge National Laboratory in Tennessee. The Jaguar can clock 1.759 petaflops and is built using 37,376 AMD processors.

The Tianhe-1A is interesting because it combines CPU and GPUs — much like desktop PCs — to create the world’s most powerful machine. In fact, Nvidia, claims if its GPUs weren’t used, then it would have taken 50,000 CPUs and twice as much floor space to create a comparable computer.

The Tianhe-1A was designed by the National University of Defense Technology in China and will be operated as an open access system for large scientific computations.

The use of GPUs in high performance computing is on the rise. Once seen in PCs used largely for multimedia and gaming, GPU-based computing has become more popular among researchers for its ability to offer raw computing power. While CPU are critical to a PC for their ability to interact with the different computing elements such as memory and disk drives, GPUs can perform specialized tasks especially related to graphics and visual computations that working in tandem with CPUs can speed up computing.

The CPU-GPU combination also helps keep up power efficiency. The system consumes 4.04 megawatts, three times less than what it would have if it were built entirely with CPUs, claims Nvidia.

Overall, the Tianhe-1A is a computing monster. It has 262 terabytes of memory and is housed in 140 refrigerator-sized cabinets.

See Also:

• Sony’s PS3 Update Could Affect Supercomputer Users
• Android Phones Can Substitute for Supercomputers
• How to Upgrade a Supercomputer, 37376 Chips at a Time
• Toshiba Crams Supercomputer Guts Into 3-D, Web TV
• Microsoft, Cray Unleash $25K Supercomputer
• Water-Cooled Supercomputer Doubles as Dorm Space Heater

Photo: Nvidia

2011 can’t come soon enough, particularly if you’re AMD. The company has been attempting to maintain hype behind its CPU / GPU hybrid since last century, but the newest demonstration of its Llano Accelerated Processing Unit (APU) is getting us back in the mood. For those who haven’t followed the play-by-play, this here all-in-one chip is expected to hit laptops and desktops at some point in 2011, and the company’s most recent demo featured a single chip simultaneously processing three separate compute-and graphics-intensive workloads. Llano was able to calculate the value of Pi to 32 million decimal places, decode HD video from a Blu-ray film and handle some other mysterious task that we may or may not ever truly find out about. Head on past the break for a promotional / informational video, and be sure to re-circle next year on your calendar. That’ll be the one… just like your high school football team always thought.

Continue reading AMD demos next-gen Llano Fusion APU, promises consumer availability in 2011

AMD demos next-gen Llano Fusion APU, promises consumer availability in 2011 originally appeared on Engadget on Wed, 20 Oct 2010 02:17:00 EDT. Please see our terms for use of feeds.

Permalink   |  AMD  | Email this | Comments

The attentive newshounds over at AnandTech have picked out a golden nugget of disclosure from AMD’s earnings call last night, citing CEO Dirk Meyer as saying “we will be launching our second-generation DX11 graphics offerings next week.” What he’s talking about, of course, is the highly anticipated refresh of the Radeon HD 5000 series, which some recent spec leaks suggested would bring a healthy bump in performance. After announcing the new product line next week, AMD promises to flood the market with “hundreds of thousands of units,” which will be shipping before the end of this quarter — meaning you’ll have a Radeon HD 6xxx in time for the holiday gaming craze if you really want it. Can’t ask anymore than that, now can we?

AMD launching next generation of Radeon graphics cards next week, shipping by end of the year originally appeared on Engadget on Fri, 15 Oct 2010 02:56:00 EDT. Please see our terms for use of feeds.

Permalink AnandTech  |  AMD  | Email this | Comments

The graphics card that doesn’t require a fridge-sized cooler is turning into something of a rarity nowadays, but we doubt the market for quiet, efficient, and halfway-decent GPUs is ever going to disappear completely. NVIDIA is fleshing out its Fermi family today with a creature that aspires to such epithets, the 96 CUDA core-equipped GT 430. It’s a patently humble GPU, as indicated by its $79.99 typical price, 49W TDP, 5.7-inch board length, and single-slot cooler design. Mind you, while those are typically considered positives, they do limit gaming performance quite significantly, with the GT 430 getting roundly beaten by ATI’s (sob!) AMD’s similarly priced Radeon HD 5670. So what niche is left for this card? Well, it’s an upgrade over integrated graphics and it gets you on the 3D bandwagon, but on the whole we’re left scratching our voluminous craniums as to why anyone would dodge AMD’s more accomplished hardware for NVIDIA’s latest. Hit up the reviews below and form your own opinion, if our one doesn’t suit your outfit today.

Read – HardOCP
Read – AnandTech
Read – PC Perspective
Read – Hot Hardware
Read – Legit Reviews
Read – Hexus

NVIDIA launches sub-$80 GeForce GT 430 for single-slot cooler enthusiasts originally appeared on Engadget on Mon, 11 Oct 2010 21:56:00 EDT. Please see our terms for use of feeds.

Permalink   |   | Email this | Comments