"The number of problems you guys are able to solve as a result of deep learning is truly amazing," Huang said, addressing more than 500 guests at NVIDIA's 'best of Utah' themed CVPR bash at the Grand America Hotel. "We've dedicated our company to create a computing platform to advance your work. Our goal is to enable you to do amazing research."
Then he called up 12 teams of researchers and presented each with an NVIDIA Pioneer Awards. The awards went to those who've used NVIDIA's AI platform to support great work featured in papers accepted by CVPR and other leading academic conferences.
But the surprises didn't stop there. Twenty guests selected at random joined Huang in the the hotel's center courtyard. There he presented them with signed, limited edition NVIDIA Titan V CEO Edition GPUs, featuring NVIDIA Volta Tensor Cores and loaded with 32 gigabytes of memory.
The limited edition GPU is built on top of NVIDIA's Volta platform. It features NVIDIA's Tensor Core architecture to enable multi-precision computing. It can crank through deep learning matrix operations at 125 teraflops at FP16 precision. Or it can blast through FP64 and FP32 operations when there's a need for greater range, precision or numerical stability.
Because this isn't a retail SKU - at least not yet - NVIDIA hasn't published official specifications for the card.