AWS Launches Next-Gen GPU-Powered Compute Instances
Amazon Web Services announced the next generation of GPU-powered compute instances, focusing on use cases such as graphics-intensive applications or the deployment of machine-learning models in production.
Amazon EC2 G4 GPU-powered Amazon Elastic Compute Cloud instances (Amazon EC2) are now generally available. These instances provide the newest NVIDIA GPUs, AWS custom Intel Cascade Lake processors, up to 100 Gbps networking throughput and up to 1.8 TB local NVMe storage.
Jeff Barr, a spokesperson, stated in a blog post that the next-generation offerings can be applied to video transcoding and game streaming applications.
“Customers with machine-learning workloads can launch G4 instances via Amazon SageMaker, AWS Deep Learning AMIs, including machine learning frameworks like TensorFlow and TensorRT, PyTorch. Caffe2, CNTK. and Chainer,” the company stated in a news release.
Barr stated that customers can now launch G4 instances in the US East, US East (Ohio), US West(Oregon), US West(N. California), Europe [Frankfurt], Europe (Ireland], Europe (London), Asia Pacific Region (Seoul) and Asia Pacific region (Tokyo). He said that they are also working to make them available in Amazon SageMaker as well as in Amazon EKS clusters.