Author(s): Raghu Gopal
Last week, the California company Nvidia unveiled a new high-performance PC graphics card, the GTX Titan X, based on the company's Pascal architecture. Capable of performing at 11 trillion floating point calculations per second, the card is one of the fastest graphics processing units (GPUs) ever developed for the consumer market.
The company said its new Titan X chip, the successor to a current model which, somewhat confusingly, has the same name, will be packaged as an add-in card for personal computers.
The Titan X card contains 12 billion transistors driving its performance. In the past, graphics cards have been used to render imagery in video games and other software as they have several specialised calculating engines rather than a few general-purpose processor cores. But recently, GPUs have been applied in a technique known as deep learning, which allows computers to teach themselves by sifting through tons of data. For example, the technique is being used to identify faces in collections of photographs and also enables cars to recognise objects in their surroundings and support autonomous driving.
The Titan X is not intended to be a mainstream card; it's aimed at serious gamers, designers and researchers looking for a high-level of image processing. The card will go on sale in early August 2016 for $1,200.
Nvidia emphasises the new GPU's ability to boost high-level virtual reality experiences, meaning the Titan X could entice early adopters of the two leading virtual reality headsets: the HTC Vive and the Oculus Rift devices. Both these headsets are tethered to a computer and require relatively powerful hardware. A solid virtual reality experience still entails a significant investment but, as is always the case, the ecosystem starts at the top and will trickle down during the coming years.
Nvidia is in a two-horse race with AMD, its main rival in the GPU market, but it also stands to see increased competition from Intel as it doubles down on this area — a long-time weakness for the other Santa Clara-based company. Nvidia is raising the bar with its newest release, showing how graphics cards will enable deep-learning networks, data-crunching systems and very powerful gaming computers capable of delivering virtual reality experiences. Nvidia is showing the world that high-performance graphics cards now mean much more than high-end gaming.