Showing items from Model Training

How GPUdirect RDMA massively speeds up machine learning at AWS

GPUdirect RDMA was invented to massively speed up the ML frameworks that so many customers use to train their models using our large fleets of NVIDIA GPUs.

Amr Ragab from EC2 Engineering took a few minutes at SC'22 to explain to us how it …

GPUdirect RDMA massively speeds up machine learning at AWS - here's why

GPUdirect RDMA was invented to massively speed up the ML frameworks that so many customers use to train their models using our large fleets of NVIDIA GPUs.

Amr Ragab from EC2 Engineering took a few minutes at SC'22 to explain to us how it …

NCCL on EFA makes the ML world go around in the cloud

Machine Learning is a huge workload, and one of the most demanding when it comes to scaling to thousands (and thousands) of CPUs. Some of the largest workloads customers run in the cloud are deep learning models, which require huge numbers …