The Amazon EC2 team were launched the first-generation Cluster GPU instances in late 2010, followed by the G2 (2013), P2 (2016), P3 (2017), G3 (2017), P3dn (2018), and G4 (2019) instances. Now, the team has launched the new EC2 P4 Instances which are powered by the latest Intel Cascade Lake processors and feature eight of the latest NVIDIA A100 Tensor Core GPUs, each connected to all of the others by NVLink and with support for NVIDIA GPUDirect.
P4 instances include 1.1 TB of system memory and 8 TB of NVME-based SSD storage that can deliver up to 16 gigabytes of read throughput per second, according to the announcement. The NVIDIA A100 GPUs, support for NVIDIA GPUDirect, 400 Gbps networking, the petabit-scale network fabric, and has access to AWS services like S3, Amazon FSx for Lustre, and AWS ParallelCluster. EC2 UltraClusters of P4d instances democratize access to supercomputing-class performance for everyday ML and HPC developers.