Nvidia Unveils The Massive Telsa V100 GPU
Nvidia has unveiled the first chip in their upcoming Volta line; the professional grade Telsa V100 reportedly boasted an R&D budget of three billion. This GPU isn’t meant for gaming, but rather engineering and AI work. In fact, Jensen Huang, founder and chief executive officer of Nvidia spoke about the importance of AI when he unveiled the chip. He said that AI (artificial intelligence) is the driving force behind great tech advances in history.
Nvidia’s seventh-generation is built around 21 billion transistors and will feature more than 5,000 CUDA cores. The chip itself was made using a 12-nanometer process more advanced than anything used in Nvidia’s current generation of GPUs. For sake of comparison, the company’s flagship Pascal GPU, the 14nm Tesla P100, offers 3,840 cores and 15 billion transistors.
The company went on to brag that the Telsa’s is streaming microprocessor architecture more efficient than the current Pascal architecture. In terms of memory, the Telsa will make use of 16GB of ultra-fast, 4096-bit high-bandwidth memory. It is unclear whether the consumer version of the Volta chips will use high-bandwidth memory or not. AMD’s Vega line will, but the technology is still rather expensive so Nvidia might go for something a bit more affordable for its consumer line of products.
The Telsa 100 isn’t just absurdly powerful and expensive to make, it’s also huge. The chip is 815 square millimeters as opposed to the Pascal’s size of 600 MM.
Obviously, this chip isn’t meant for consumers or gaming, but it does give us a glimpse into what Nvidia is planning for the upcoming Volta line.
Eric is an avid tech junkie, gamer, and comic fan. When he's not working on his PC, you'll find him at your local comic book shop.