Issue link: https://read.uberflip.com/i/1187300
Amazon Web Services – HPC Lens AWS Well-Architected Framework Page 31 Performance Efficiency Pillar The performance efficiency pillar focuses on the efficient use of computing resources to meet requirements and on maintaining that efficiency as demand changes and technologies evolve. Design Principles When designing for HPC in the cloud, a number of principles can help you achieve performance efficiency: • Design the cluster for the application: Traditional clusters are static and require that the application be designed for the cluster. AWS offers the capability to design the cluster for the application. A one-size- fits-all model is no longer needed. When running a variety of applications on AWS, a variety of architectures can be used to meet each application's demands. • Test performance with a meaningful use case: The best method to gauge an HPC application's performance on a particular architecture is to run a meaningful demonstration of the application itself. An inadvertently small or large demonstration case–one without the expected compute, memory, data transfer, or network traffic–will not provide a meaningful test of application performance on AWS. Although system-specific benchmarks offer an understanding of the underlying compute infrastructure performance, they frequently do not reflect how an application will perform in the aggregate. The AWS pay-as-you-go model makes a proof-of-concept quick and cost-effective. • Consider cloud-native architectures: In the cloud, managed, serverless, and cloud-native architectures remove the need for you to run and maintain servers to carry out traditional compute activities. Cloud allows each step in the workload process to be decoupled and optimized. • Experiment more often: With virtual and automatable resources, you can quickly carry out comparative testing using different types of instances, storage, or configurations. Definition There are four best practice areas for performance efficiency in the cloud: