AWS announced general availability of Amazon EC2 Trn1 instance supplied by AWS training chips. Trn1 instances offer the best performance for deep learning training of popular machine learning models on AWS, while saving up to 50% on training costs over comparable GPU-based instances.
AWS training A second-generation machine learning chip built by AWS for deep learning training. Each Amazon EC2 Trn1 instance deploys up to 16 AWS Trainium accelerators to provide a high-performance, low-cost solution for deep learning training in the cloud.
The Trn1 instance is a second generation AWS elastic fabric adapter A network interface for better scaling efficiency.
Also use Trn1 instance AWS Neuron, the SDK for Trn1 instances. This allows customers to get started with minimal code changes and integrates into popular machine learning frameworks such as: PyTorch When TensorFlow.
The Trn1 instance is AWS Nitro Systemis a collection of AWS-designed hardware and software innovations that streamline the delivery of isolated multi-tenancy, private networking, and fast local storage.
According to AWS, developers can run deep learning training workloads on Trn1 instances. AWS Deep Learning AMI, AWS Deep Learning Containeror managed services, etc. Amazon Elastic Container ServiceWhen AWS Parallel Clusterwith support Amazon Elastic Kubernetes Service, Amazon SageMakerand so on AWS Batch.
As computer-intensive workloads increase, the need for highly efficient chips increases dramatically. Trainium can be compared, Google’s Tensor Processing Unit These are AI training workloads hosted on Google Cloud Platform, and the services offered vary on many levels.
It can also compete with some of the newly launched AI chips. IBM Power10Claims to be 3x more efficient than previous models in the POWER CPU series, or NVIDIA A100which claims to offer 6x better performance than NVIDIA’s previous-generation chips.
