Nvidia's $2999 Titan V is a 110 teraflop GPU for AI processing

Adjust Comment Print

As you might expect, the TITAN V will be added to the NVIDIA GPU Cloud, which provides quick access to a fully optimized deep learning software stack to advance artificial intelligence (AI) development.

The Titan V promises 110 teraflops of raw computing capability - nine times that offered by its predecessor, the Pascal based Titan X - whilst consuming the same amount of power. Nvidia has pointed towards a redesign of the streaming multiprocessor that is at the centre of the GPU, which it claims doubles energy efficiency compared to the previous generation, which results in the boost in performance in the same power envelope.

"Our vision for Volta was to push the outer limits of high performance computing and AI. We broke new ground with its new processor architecture, instructions, numerical formats, memory architecture and processor links", Jen Huang, Nvidia founder and CEO, said in a statement.

All of that power is packed into the 21.1 billion-transistor GV100 GPU that is manufactured using a TSMC 12nm FFN high-performance process, customized for NVIDIA.

More news: Alert on oversold position :Edison International (EIX) stock RSI moves to 17.17
More news: West Brom boss Pardew fears Evans departure in January
More news: Facebook Now Allows You to Live Stream While Playing Games

Nvidia has unleashed a new Titan graphics card which it's billing as the most powerful GPU in the world.

Well, it's possible. I already mentioned that I did not see an SLI connectors on the card, and made a comment about NVLINK, indeed the card has NVLINK Bridge connectors, which is a fundamental change, but also price.

The Titan V has 12GB of stacked HBM2 memory which operates at 1.7Gbps on a 3072-bit bus, for a total bandwidth capacity of 652.8GBps.

While the Pascal architecture that powers Nvidia's GeForce 10-series gaming GPUs is far from showing any signs of age, the company has just released its first consumer product based on the next-gen Volta architecture. NVIDIA TITAN V maxes out the GV100 silicon, if not its memory interface, featuring a whopping 5,120 CUDA cores, 640 Tensor cores (specialized units that accelerate neural-net building/training). The GTX 1080 still remains the best gaming GPUs available for the consumer.