Titan X is the next-gen graphics processor which powered the much-buzzed-about Oculus Rift virtual reality demo produced by Epic Games and Weta Digital at the Game Developers Conference in San Francisco a couple of weeks ago, Nvidia CEO Jen-Hsun Huang said. Called "Thief in the Shadows," the immersive VR experience put viewers face-to-face with Smaug the dragon from Peter Jackson's The Hobbit movies, knocking just about every reviewers' socks off. Now the powerful graphics card that helped make "Thief in the Shadows" such a hit can by yours for $999.
These days, you can buy an entire desktop PC for half that, of course. But with Titan X, you're getting 3,072 graphics processing cores which deliver 7 teraflops of peak single-precision performance, Nvidia said. Try to get that kind horsepower out of the $499 rig you just picked up at Costco, partner.
Titan X has 12GB of onboard memory and uses it very, very fast with 336.5GB/s of memory bandwidth. The Maxwell-class GPU "delivers twice the performance and double the power efficiency" of the GeForce GTX 980, according to the graphics chip maker.
"The latest AAA titles are breathtaking on Titan X in 4K. Middle-Earth: Shadow of Mordor, for example, runs at 40 frames per second on high settings with FXAA enabled, compared with 30fps on the GeForce GTX 980, released in September," Nvidia said.
Jen-Hsun, keynoting the first day of GTC in San Jose, Calif., also stressed the capabilities of the new GPU for scientific and research applications.
It's been almost a decade since Nvidia began consciously pivoting away from its reputation as company that only makes products for PC gaming enthusiasts. The company released CUDA eight years ago and sells bucketloads of GPUs to builders of HPC and supercomputing systems these days.
With Titan X, Nvidia is pushing hard into "deep learning," a flavor of artificial intelligence development that the company believes will be "an engine of computing innovation for areas as diverse as advanced medical and pharmaceutical research to fully autonomous, self-driving cars."
Nvidia's new flagship GeForce GPU has deep-learning capabilities baked right in and Nvidia on Tuesday also introduced two complementary software and hardware development platforms, dubbed DIGITS, to help developers get the most out of those features in Titan X.
The new DIGITS Deep Learning GPU Training System software is an "all-in-one graphical system for designing, training, and validating deep neural networks for image classification," Nvidia said. The DIGITS DevBox is the "world's fastest desk-side deep learning machine" with four Titan X GPUs at the heart of a platform designed to accelerate deep learning research.
Nvidia also offered some insight into its future product roadmap, detailing how its next-gen GPU architecture code named Pascal will further incorporate machine-learning functionality.
"Nvidia's opening keynote at GTC 2015 was all about showing how serious the company is about deep learning as it pertains to neural networks and machine learning," said Patrick Moorhead, principal analyst for Moor Insights & Strategy. "Their Titan X, DIGITS DevBox and Pascal roadmap update all point towards the company moving towards a much more serious direction in this field.
"Deep learning is the next logical step for the company in trying to improve the world of visual computing while also driving demand for GPUs and their GPU technologies."
Source
No comments:
Post a Comment