What’s new with our EPIC on-demand HPC tool?

EPIC HPC

To ensure that EPIC, our on-demand simulation tool, offers an effective solution for users, we’re continually updating it, adding innovative features, the latest HPC resources, and new applications.

Here are just some of the highlights from the past few months that are already bringing benefits to our community of users.

One-off payments / Pay as you go HPC 

We have added a new payment method to give our users more flexibility when paying for HPC on demand. Previously, users had the option of paying via a credit card on a monthly invoice or using a monthly BACS invoice. 

Now, a pre-pay option has been added, allowing you to top up your account using a credit card, as and when you wish. This new solution means no monthly commitment and will enable you to test out the platform effortlessly, without worrying about any unexpected charges.

EPIC’s payment options have also all been updated to comply with the new SCA payment requirements, giving you additional security for any payments made using a card online. 

Post-process in the cloud with EPIC

How do we solve the issue of post-processing large datasets? Running HPC jobs often leads to huge volumes of data being generated. With EPIC, users already have the option to download this data to a local machine and post-process it, or launch a ParaView server remotely. However, sometimes, you need to be able to combine tools and scripts to extract the data you require. 

EPIC now offers you the ability to start interactive Linux Desktop instances to post-process your results, combining the flexibility of accessing the data on a desktop machine without the need to download enormous datasets. The Nice DCV provides very high-performance GPU accelerated interactive nodes that you can start and stop via EPIC in a matter of moments.  

EPIC also now lets you mount your EPIC datastore on the interactive node, meaning all of your simulation data is available. What’s more, any new files or changes you make to your data are automatically synchronized with EPIC.

Run your OpenFOAM* solver on AWS

EPIC offers a range of HPC backend systems, both traditional supercomputing and cloud platforms. This mixed offering allows our users to choose a balance between performance and cost that suits them. Recently, we have included in EPIC, some new compute node options from AWS, including the new C5 Bare Metal instance type.

By adding these instance types to EPIC, we enable our users to run HPC jobs on them with zero changes; they select the new node types during job submission, and that’s it. The new node types have been particularly popular with our OpenFOAM users, with the AWS spot pricing giving them access to large scale compute resources at a very competitive price. 

 Interactive OpenFOAM node based on DCV

The batch job features of EPIC allow you to run the solver component of your OpenFoam workflow on a variety of HPC providers. This solution is an excellent fit for those looking to run the solver at scale, but sometimes you need to run more than just the solver. 

To enable more flexible use cases, we have created a new interactive OpenFoam node based on the NICE DCV infrastructure.  This new feature allows you to start an OpenFOAM desktop on-demand and access a high-performance cloud-based desktop via your browser.

The desktop can have access to all of your EPIC data and comes with a variety of hardware configuration options.  The nodes currently have several pre-compiled and configured OpenFOAM versions available, including V1906, V7.0 and Foam Extend 4.0. Paraview is also installed and ready to use. These nodes are perfect for case setup, meshing, solving and post-processing.

Benefit from EPIC today

If you are interested in using EPIC or would like to test out the new features, click the link to find out more.

*This offering is not approved or endorsed by ESI Group or ESI-OpenCFD®, the producer of the OpenFOAM® software and owner of the OpenFOAM® trademark.

 

 

 

0 Comments

Leave a reply

Your email address will not be published. Required fields are marked *

*