Loading Events

« All Events

  • This event has passed.

Webinar: Workload Dependencies: Lift-and-shift Successfully to HPC in the Cloud

October 10, 2023 @ 8:00 AM - 9:00 AM

HPC Digital Webinar CloudComputing 700x500 02

Optimising HPC in the Cloud

Strategies for Successful Innovation

There are a multitude of good reasons to migrate your HPC workloads to the cloud.  Whether you’ve made the decision to test run your HPC workloads in the cloud or to immediately move production workloads to the cloud, there are several finer points to consider to ensure a successful outcome of your decision.


Confidently into the cloud

Altair HPC experts will discuss in this webinar series a couple of key strategic decisions and preparation steps that you should undertake before taking your first steps to the cloud:

  • Multi-cloud HPC deployments: Keep your options open to select appropriate cloud HPC suppliers depending on various factors, incl. workload type, pricing, location of data centres, etc.
  • Automated cloud-bursting strategies: Spinning compute nodes up/down depending on business requirements and restrictions offer cost optimisation potential. Understanding strategies for automating this process is key to cost-efficient HPC in the cloud.
  • Workload dependency planning: Ensuring that you have everything that you need to run complex workloads in the cloud, potentially with several dependencies on data or software library availability, is critical to fast and efficient workload execution.


Workload Dependencies: Lift-and-shift Successfully to HPC in the Cloud

Rosemary Francis,
October 10, 2023
14:00 CEST | 8:00 EDT | 21:00 JST

Moving HPC workloads to the cloud with high confidence of success requires careful planning of which data and software resources are required to run the workload and how to make them available in the cloud. This session offers guidelines on how to audit and characterise workload dependencies to ensure that all requirements are provided for when moving workload to the cloud.


Share this post via: