After clicking “Watch Now” you will be prompted to login or join.
Click “Watch Now” to login or join the NVIDIA Developer Program.
Accelerating Distributed DL Training with Optimized GPU Computing Solution
Andy Kuo , Super Micro Computer, Inc. | Carlos Weissenberg, Super Micro Computer Inc.
High-performance multi-GPU computing plays an important role to accelerate distributed Deep Learning training. Choosing the right GPU Topology could impose a considerable impact on training efficiency. This section mainly addresses key factors of distributed Deep Learning training, the hardware technologies as well as the different GPU based hardware solutions.