Many engineering companies facing tight equipment budgets or needing a quick boost in horsepower find the idea of a pay-as-you go cloud service enticing. But what’s really in it for you? Not all that stuff about freeing up workstations to do other things while a job runs on the cloud, but what’s it like feeding and running a CFD (computational fluid dynamics) job on one of these rigs? How do you set up and execute a job? How does its workflow flow? How do you know the cloud outfit has its security act together? What are the questions you should be asking providers?
Whether for a one-off for a big project, a tool to get you through a crunch or as a strategic alternative to an internal HPC (high-performance computing) infrastructure, the cloud service option can confront you with a ton of questions. Today’s Check it Out link takes you to an on-demand webinar that offers answers to many of your questions as well as a few that you might not have come up with yet.
“Cloud Computing and Workflow with Rescale” is a joint presentation from Convergent Science, the developers of CONVERGE CFD software, and Rescale, a cloud platform tuned for simulation. What you have here is as straightforward a presentation as you’re going to find when it comes to showing you how you interact with a cloud service.
This 45-minute webinar begins with a brief overview of Convergent Science and what its CONVERGE CFD software is all about. CONVERGE merits your attention for a lot of reasons, but chief among them is that it dispenses with manual grid generation. It automatically generates meshes at run-time for each time-step. Hands-on, that means you import a model and clean it, flag boundaries, set up parameters then solve. That’s the workflow, and it operates pretty much the same way with Rescale.
The Rescale presentation also begins with the who-we-are overview. Simply put, this cloud service is for engineering outfits whether they need a one-time boost in computing capacity or the horses to run massive jobs, say, design of experiments, without tying up resources. Its sandbox is CFD, FEA (finite element analysis), molecular dynamics, NVH (noise, vibration, and harshness) or whatever your compute-intensive engineering job. Rescale partners with leading CAE developers such as Convergent Science to ensure that its hardware is optimized for your CAE software of choice.
What’s striking in the first part of the presentation are the insights it provides. If you keep careful notes, you will build an excellent checklist for vetting a cloud service. For example, do you know if your software has an on-demand option? Can you understand and use the cloud provider’s interface? How does the provider ensure the privacy and security of your data? Does a third-party vouch for what it says about security?
Still, the highlight of Rescale from a user’s angle probably will be its workflow tools, and that’s what you’ll see in operation for a majority of this webinar. Basically, the Rescale workflow goes like this: You upload a CONVERGE file to Rescale, set up the job then run it. The demo shows you how you do it, including building a private cluster, entering analysis parameters, setting conditions and configuring and managing hardware resources like cores, resolution and so forth. (Hint: Hit the full-screen mode and clearly see what’s going on during the demo.)
That’s pretty much its basic workflow. You can invite in collaborators, get help from Rescale experts on setting up a job, monitor results even from a mobile device and clone a job to use later.
CONVERGE on the Rescale cloud seems like what such a gig should be. Hit today’s Check it Out link to watch the “Cloud Computing and Workflow with Rescale” webinar to see for yourself what this partnership could do for you. If you like what you see, a complimentary trial of CONVERGE on the cloud is readily available. Good stuff.
Thanks, Pal. – Lockwood
Anthony J. Lockwood
Editor at Large, Desktop Engineering