Client Overview 

Achates Power is the developer of radically improved internal combustion engines that increase fuel efficiency, reduce greenhouse gas emissions and cost less than conventional engines. Achates Power Opposed-Piston Engine allows global original equipment manufacturers (OEMs) to achieve the world’s most stringent current and future fuel efficiency and emissions standards without additional cost or complexity.

Challenges

Achates Power needed additional compute capacity to meet a short term high priority project. The standard delivery time for new equipment to add to their existing HPC cluster would take too long for the project, immediate on-demand compute power was needed that engineers could quickly utilize with minimal impact.

Solution

“TotalCAE made it simple to incorporate burst to cloud with our existing TotalCAE cluster. We just pointed our browser to a new URL and we were able to keep using the same intuitive web interface to run our applications on AWS.”

– Rodrigo - Achates Power Senior Engineer

The TotalCAE Public Cloud on AWS solution was implemented to achieve the ability to to double engineering throughput.

TotalCAE Public Cloud for Engineers is an easy to use HPC Cloud on AWS Managed by TotalCAE. The entire solution is fully managed remotely by TotalCAE, alleviating any need to worry about how to maintain, manage, and update your HPC cloud.

TotalCAE on AWS offers clients several benefits. 

This case study will help you understand how TotalCAE enables engineers like at Achates Power to reduce simulation runtime on AWS, increase their engineering agility, and maximize their investment in engineering staff, CAE licensing, and AWS resources.

TotalCAE on AWS - Three Clicks to Solve

Running TotalCAE on AWS is very simple. You point your browser to a persistent URL set up for your company, and you are presented with a simple to use web interface shown below to submit your engineering applications to the cloud.

The three steps to submit are:

Just three clicks to solve. There is no complicated setup or wizards to click through so you can focus on engineering and less on job setup. To see a video of using our cloud platform visit https://totalcae.com/aws  For power users that prefer the command line, full terminal access is provided along with our command line submission tool, called tsubmit. 

After you submit your job, the TotalCAE scheduler will automatically power on the appropriate compute power, and run your job. The nodes will be powered off when your job completes to save money as the compute resources are “pay as you go”. 

TotalCAE makes it easy to submit hundreds of supported applications to AWS Cloud Infrastructure listed at https://www.totalcae.com/engineering-applications  

Job Cost and Billing Management

When utilizing pay as you go resources, it is important to give engineers the tools they need to understand and control their simulation costs. The TotalCAE portal gives you an estimate of the on-demand job cost so the engineer can gauge how much a particular simulation job run will cost. 

The cost of a particular job will depends on the type of hardware, and how many CPUS you need to utilize. Dialing up and down the CPUS will show you the job cost impact.

Engineers can cap a particular job cost by selecting a maximum runtime, which will give you the maximum job cost in the event the simulation does not converge, or has other errors, or if the solution isn't worth it if it exceed the max spend.

Finally users can select amongst various projects they have access to, for billing the particular job against. This is useful to make sure that simulation dollars allocated for various projects stay in budget, and to help engineers control spending based on simulation budgets. This data is then available in reports that can be viewed on-demand for charge back, auditing, or project cost review.

Avoiding Download Wait Issues

Engineers are often concerned about the bandwidth challenges with respect to downloading  and uploading hundreds of gigabytes of CAE when you have a very interactive workflow. Many CAE applications generate a large amount of simulation output data. TotalCAE offers several solutions on AWS to overcome this challenge.

Avoid Waiting on Download - Post Process on the Cloud

One solution to the download issue is to avoid the downloading every result and do the majority of the post processing on the cloud. 

TotalCAE supports cloud desktops our remote visualization technology, which is a high-performance remote display protocol that enables you to do post processing on the cloud for most popular graphical CAE applications. In this way, you can look at your results, do post-processing, and bring back only the final results you may need to preserve back on-premise.

Avoid Waiting on Download - TotalCAE Automatic Download While you Sleep

Even with cloud desktops, engineers will want to bring back some result data from their AWS infrastructure back to on-premise.  TotalCAE has an auto-download tool that will start streaming the job results back on-premise automatically when the job completes, so the user does not have to click "Download" and then wait.

For example, if your simulation finishes on the cloud at midnight while you are asleep, the results automatically start streaming back on-premise so that when you arrive to work in the morning, the results are downloaded and waiting for you to post process with no time was wasted waiting for a download to complete.

Avoid Waiting on Download - AWS Direct Connect

Your network IT team can setup AWS Direct Connect which is dedicated connection to transfer data to and from AWS directly. Network latency over the Internet can vary given that the Internet is constantly changing how data gets from your company to AWS. With an AWS Direct Connect dedicated connection you will get more consistent transfer times over Internet-based connections.

Monitoring Simulation Jobs

Pay as you go jobs running on AWS are being billed by the hour by TotalCAE, and engineers often can't monitor their jobs every hour of the day to make sure the solution is converging, or that the model isn't having issues. This can results in getting larger bills for jobs having issues, or requiring the engineer to constantly monitor their job to avoid getting a bill for a job that may not deliver a result.

TotalCAE reduces this issue by immediately notifying the engineer when something is looking amiss with the model, or various application specific issues so the engineer can decide to cancel the job, or take some remediation.

For example, if the application  solver output has not updated but the solver is still running, or if solver appears running but is queued for licensing, and various other solver specific issues can be detected and alert you via the portal and email to avoid you having to constantly check in on the run to see if it is iterating correctly to avoid over spending.

If you prefer to monitor the job output yourself, you can still tail the output manually.

Accessing CAE Application Licensing 

The primary method of accessing CAE application licensing on AWS is to bring your own license, where TotalCAE will forward your on-premise licensing from your on-premise license server to AWS to utilize your existing licenses. 

TotalCAE also will host your license server in AWS, to enable you to access them easily from on-premise or the cloud to enable you to use your licenses anywhere.  

Some vendors, like SIEMENS STAR-CCM+ have a cloud hosted option where you purchase a Power on Demand (PoD) key from SIEMENS, and just copy and paste that key into the TotalCAE web portal to utilize your PoD key.