Jen-Hsun Huang, CEO of NVIDIA Inc., finally announced the technical details of the much-talked-about high-power computing GPU, the GP100, during the opening keynote of the GPU Technology Conference or GTC 2016. This is the first product in a series that the company will soon be releasing, and it will be equipped with the latest Pascal technology. The company will be ready to ship it by June while OEMs and consumers can get their hands on the monstrous GPU by early 2017.
The GP100 GPU is based on the Tesla P100 board, which uses the latest 16nm process. This means that NVIDIA can squeeze twice the number of transistors into the same amount of space used by the previous technology, the 28 nm process. It is 15.3 billion transistors against 8 billion, and this new technology can offer a great amount of increase in computing power as well as being more power efficient.
PCWorld reported that the Pascal GP100 GPU can accommodate 3840 CUDA cores and 224 texture units into 64 streaming multiprocesors (SM) in a 610 mm<sup>2 size chip. By comparison, the previous Titan X and Tesla M40 using Maxwell technology contain 3072 CUDA cores. The L2 cache for the Pascal-based board is 4096 KB HBM2, while the old one offers 3072 KB GDDR5.
Huang spent two hours explaining how the new NVIDIA Pascal-based GPUs will benefit supercomputers/deep learning, AI acceleration computing, data centers and cloud computing servers. The company has created the most advanced GPU architecture that the computing world has seen so far. Researchers will be able to solve more complex and larger problems in the fields of cosmology, seismology, climatology, material science, and a whole lot more, as WCCFTech reported.
NVIDIA also announced that their new Tesla GPU accelerators will be used when the Piz Daint system receives its upgrade. The Piz Daint is Europe's fastest supercomputer and is located at the Swiss National Supercomputing Center (CSCS) in Lugano, Switzerland. It is currently ranked number seven among the Top 500 list of supercomputers in the world.
Ian Buck, vice president of Accelerated Computing at NVIDIA, said, "CSCS scientists are using Piz Daint to tackle some of the most important computational challenges of our day, like modeling the human brain and uncovering new insights into the origins of the universe."
He then added, "Tesla GPUs deliver a massive leap in application performance, allowing CSCS to push the limits of scientific discovery."
The gaming community, who are also looking forward to experiencing the computing power of the new Pascal technology, should not be disappointed. Consumer variants using the latest technology are expected during the ongoing GTC 2016.
