Computer Systems and Engineering Seminar Series
Inside Pascal: NVIDIA's Newest Computing Platform
||Friday, September 16, 2016
||11:45am - 1:00pm
||North 311, Duke
||Light lunch will be served.
Lars Nyland will present details of the Tesla P100 and the Pascal GP100 architecture. The GP100 GPU used in Tesla P100 incorporates multiple revolutionary new features and unprecedented performance. Key features of Tesla P100 include:
Extreme performance - powering HPC, deep learning, and many more GPU Computing areas:
NVLink - NVIDIA's new high speed, high bandwidth interconnect for maximum supplication scalability;
HBM2 - Fastest, high capacity, extremely efficient stacked GPU memory architecture:
Unified Memory and Compute Preemption - significantly improved programming model;
16nm FinFet - enables more features, higher performance, and improved power efficiency. To learn more, read the Tesla P100 white paper.
Lars Nyland is a senior architect at NVIDIA, where he has worked in the compute architecture group for 10 years. His work covers all aspects of architecture related to HPC computing on GPUs, from floating-point issues to memory performance. Prior to joining NVIDIA, he received his PhD at Duke University and was a professor at the Colorado School of Mines and UNC-CH.
Hosted by: Alvy Lebeck