A Guide to High Performance Computing in the AWS Cloud

Print Friendly, PDF & Email

In the world of HPC, on-premises infrastructure is increasingly falling short when it comes to speed and capacity for these complex workloads.

AWS cloud

Download the full report.

Often, an enterprises’ top engineers and researchers have to compete for access to run their HPC workloads on limited, on-premises capacity, which can cause bottlenecks and long queues.

Further, hardware updates often take months to get up and running. A new ebook from Intel and AWS explores how AWS cloud can mitigate some of these challenges and enable engineers and researchers to “innovate without constraints,” as well as allow for greater scaling and agility often lacking in on-premises setups.

AWS users also have easy access to Intel optimized software libraries via the AWS marketplace.

The ebook begins by asking readers, “What could you accomplish with a million cores?” In today’s environment, increasingly large data sets and resource-intensive algorithmic processing continues to challenge the limits of HPC infrastructure. That said, most HPC processing is still conducted using on-premises systems. But for those companies unwilling to invest in new hardware to handle peak demands, moving HPC workloads to the cloud may be the answer. AWS cloud solutions with Intel® Xeon® processors can help improve compute performance, memory capacity and bandwidth and I/O scalability.

“The highly customizable computing platform and robust partner ecosystem enable your staff to imagine new approaches so they can fail forward faster, delivering more answers to more questions without the need for costly, on-premises upgrades,” the report states.

AWS cloud can also help you access resources faster, and migrating your HPC application to AWS means you’ll no longer have to “make tradeoffs between experimentation and production.” Instead, you can quickly reconfigure resources to match your workload’s needs and launch your HPC clusters in minutes for faster time to results.

Moving HPC workloads to the cloud also makes it easier for businesses to keep up with the latest technology. No more costly hardware overhauls. Users can migrate to newer, more powerful Intel® Xeon® based EC2 instances as soon as they are made available on AWS. AWS cloud solutions also work to remove some of the traditional challenges associated with on-premises clusters, including fixed infrastructure capacity, technology obsolescence and high capital expenditures.

Moving HPC workloads to the cloud also makes it easier for businesses to keep up with the latest technology.

AWS also helps enterprises to collaborate securely, as keeping your on-premise infrastructure compliant with an “alphabet soup of regulations and legislation” can be challenging. The AWS cloud is compliant with the latest revisions of GDPR, HIPAA, FISMA, FedRAMP, PCI and other regulations.

Lastly, the report walks readers through the potential economic benefits of AWS, which offers on-demand pricing for short-term projects, contract pricing for long-term, predictable needs, and spot pricing for experimental work or research groups with tight budgets.

The full ebook, available here, includes the following chapters:

  • What Could You Accomplish with a Million Cores?
  • Access Resources Quickly
  • Leverage Latest Technology
  • Collaborate Securely
  • Innovate Economically
  • Explore the Possibilities
  • Start Experimenting

Download the ebook, “Innovation Without Limit: Your Guide to High Performance Computing in the Cloud,” courtesy of AWS and Intel, to explore whether AWS cloud is the right option for your business.