Process and culture - High Performance Computing Lens
HPCSUS04-BP01 Promote a culture of constant monitoring and performance improvementResources

Process and culture

HPCSUS04: What are your methods to rapidly introduce sustainability improvements?

With the energy-intensive nature of HPC workloads, it is important to constantly monitor and continually adopt new technologies to reduce your environmental impact. You can test and adopt new technologies quicker in the cloud when compared to a traditional on-premises environment.

HPCSUS04-BP01 Promote a culture of constant monitoring and performance improvement

HPC workloads aggregate computing power and storage to allow for fast processing and calculation to solve scientific, mathematical, and engineering challenges. As a result of scale, HPC workloads are often more energy-intensive than general purpose computing. Therefore, better use of resources (both hardware and software), coupled with shorter runtimes, can lead to improved utilization and environmental sustainability.

Implementation guidance

  • Since the impact of HPC is relevant in achieving your sustainability goals, it's important to promote a culture of constant monitoring and improvement for this workload. On-premises HPC clusters usually have a life cycle that can spans years during which they are rarely updated. When running HPC in the cloud, leaders should promote a culture of continuous innovation. This helps improve performance, reduce cost, and improve sustainability.

  • To continually improve and streamline your HPC workloads, you can automate your cluster deployment using CI/CD pipelines to easily test and deploy potential performance improvements and limit errors caused by manual processes. For more information, see Tutorial: Create a simple pipeline (S3 bucket).

Resources