How did you first become involved with engineering simulation in your career?
In one of my grad school courses, I had to write a program to simulate a molecular dynamics model of thermodynamic phase change. I used a graphics program to visualize and animate the motion of the molecules in time. I was utterly mesmerized watching them jiggle themselves from solid into liquid during melting, and seeing the solid phase grow from nucleation sites during freezing. Then by playing with the molecular interaction forces, I could control the thermodynamics and phase diagram exhibited by the system – that blew my mind. From then on, I was hooked on computational modeling and simulation of all kinds.
Why is OnScale participating in the Revolution in Simulation initiative?
OnScale wants to make massive simulation power easily accessible to as many people as possible, so the synergy with Rev-Sim makes participation irresistible.
Why is expanding the use of simulation technologies through the democratization of simulation important to your customers?
Our customers are people who want to explore innovative designs and engineering ideas, who want to make things that work better, and who want to leverage the power of simulation without the usual constraining license schemes and frustrating usability problems. They recognize that simulation technology brings huge value when fully unleashed, whether it’s running design studies across many parameter variations or modeling the complete behavior of a complex device. Democratization means making that value accessible to a broad audience in a way that is time and cost efficient. It also means building in sufficient guidance and intelligence to ensure that simulation results are accurate and reliable and avoid the all-too-common garbage-in-garbage-out situation.
What do you see as the most important benefits?
The key benefits of simulation democratization are enhanced design evaluation speed, efficiency, agility, and collaboration for a team or company. Instead of just a few experts being able to use simulation tools and interpret results, with democratization, designers, engineers, testers, and managers across an organization can participate in and contribute to the knowledge and decision impact generated via simulation.
Can you share with us an example of a benefit from one of your customers?
Sure, one example that we can share publicly is with a medical ultrasound customer, Verathon. Verathon develop an ultrasonic sensor for bladder monitoring, indeed are a market leader, but were acutely aware that MEMS technologies are the future of the industry and they needed to undergo a product redesign – no mean feat given the complexity of MEMS and a short time frame to hit a prototype fabrication deadline with a foundry partner. They needed to accurately evaluate 1000s of potential designs in a two-week period, so the only possible solution was digital prototyping of those designs. Each simulation was a complex, transient, multiphysics problem, and the actual clock time to conduct all simulations serially would be at least one month (or more depending on the solver). By providing an easy to use workflow, OnScale enabled Verathon to quickly spec the design, identify the permutations to study, and run over 3000 simulations in parallel. The entire simulation process took under an hour using 8 cores per simulation executed in parallel on AWS. The workflow further allowed Verathon to quickly assess the KPIs from the resulting data and settle on the optimal design, with time to spare before tape-out (finalized design ready for fabrication).
What are some of the challenges that you have also witnessed?
Some folks still struggle to understand the value of simulation, viewing it as unreliable or unnecessary. Sometimes it is complicated to add simulation into the overall development workflow in a way that is not disruptive. Another common challenge is generating a computational mesh based on the input CAD geometry when the CAD designer was not aware of the requirements for using CAD as input for simulation – unnecessary details may be present, or edges that are supposed to be coincident are not perfectly aligned (for example). Such issues tend to confound the typical meshing process and require involvement of an expert.
What does OnScale provide to help industry overcome these hurdles?
OnScale strives to make complex, powerful simulation as easy as possible to access and use, removing all the barriers for design and engineering teams to quickly experience the value of simulation for themselves. With our consumption based pricing (pay-as-you-simulate), it becomes cost effective to test out simulation across a range of test cases in order to evaluate its ROI. For complex workflows involving many tools, OnScale has an API that allows simulation to be included within scripted workflows. As for CAD, though OnScale is not a CAD tool and cannot directly fix CAD issues, OnScale users are alerted to problems before they try to run simulations, and OnScale is developing technology to make our simulations fault-tolerant to many common problems related to CAD geometry.
Where can readers go to learn more?
At www.onscale.com we have a wealth of demo’s, simulation guides, white papers, and other material about the OnScale technology and its applications.
Are there any other questions you would like to address?
We hope readers will imagine what it means for the pace of technology development to democratize access to supercomputers for engineering simulation. OnScale plans to further build on that foundation to democratize the use of massive data sets of simulation results (“synthetic data”) for machine learning applications. We believe this will have a profound impact on the next generation of engineered products and devices.
Thank you Dr. Freed! You are invited to connect with David on LinkedIn at https://www.linkedin.com/in/davidmfreed/.