High performance computing (HPC) refers to the harnessing the processing power of thousands of CPUs running in parallel together with associated networking, memory and storage to solve complex, data-intensive tasks.
High performance workloads need to be tightly bound with high bandwidth and application, storage and compute all in the same place to ensure low latency. The capital outlay for HPC can be very large, with some environments costing as much as £100 million to set up on-premise.
With such high costs, IT teams are charged with making it pay, challenged to deliver very high utilisation rates, typically over 90 per cent. Workloads may take days to run and IT teams must schedule effectively, and keep line of business stakeholders happy - ensuring the wait for critical data is minimised and avoiding frustration from potential delays.
Enterprises commonly use a mixture of on-premise and cloud, managing what they can in house and bursting into the cloud when demand from the business exceeds capacity.
Join us on Tuesday 1st September at 11am for a live web seminar during which our expert panel will discuss how enterprises can rise to the financial and utilisation challenges of HPC, make their infrastructure fit for the future, and achieve the best possible performance in the cloud. We will reveal the results of exclusive research into key HPC trends:
- How enterprises are making use of cloud and on-premises infrastructure for HPC
- Which hyperscale cloud providers are best for HPC
- What drives choice of cloud HPC providers
- Satisfaction with providers
- Key challenges and frustrations with HPC
Register today and join us to find out how UK enterprises are optimising utilisation, keeping business stakeholders happy and aligning HPC with business strategy.
Keep it simple, advises David Keigher
We speak to Canonical, Red Hat and SUSE about the place of Linux in a cloud-based future - and what the CentOS EOL foretells
There’s a lot of sensitive data contained in Office documents - so it makes sense to take care of it
'We needed a much more flexible and cost-efficient operation to keep up with the big players in the field', says head of DevOps Chris Callaghan