CD-adapco’s Scalable CFD Technology Now Available on the Microsoft Azure Cloud

How HPC Scalability for Simulation Creates Digital Prototypes of Big Scale Projects

Scalability of STAR-CCM+ on Microsoft Azure using the “Le Mans 100 million cell” test model. (Image courtesy of Microsoft Azure.)
STAR-CCM+ has now become available on the Microsoft Azure cloud. With a high-speed, low-latency network fibre that uses RDMA and Infiniband technology, Microsoft Azure allows for scalability of STAR-CCM+ models.

With this development, detailed high-fidelity models are now readily available to the engineering community. This allows for the most complex of engineering designs to be scrutinized and optimized with reduced time costs.

In addition to reducing computational time, STAR-CCM+ also offers Power Session, Power on Demand and Power Token licensing options. These combined licenses allow the user to select what suits them most to get the most out of their investment. 

 

Simulation versus Prototyping: Big Scale Projects

The pursuit of optimised technologies often brings us to the question: “Should I make and test a prototype or simulate the system?”

For small scale testing, either method is acceptable, but suppose you want to test the non-linear behaviour of a wind turbine blade or the acoustic field behind a jet engine. In cases like these, full-scale testing is no longer an option.

Large-scale testing facilities have limited availability. Severe operating environments can prevent the collection of raw data. In addition, the time and financial cost associated with physical testing can be sufficient to rule it out as an option.

The Challenges to Simulate Big Scale Projects

Considering the above, numerical simulation of these test cases may be the most effective solution, especially when time and financial factors are considered.

However, this has only recently become the case. For years, computational resources have been playing “catch-up” with numerical modelling codes. So much so that early research studies using numerical methods were highly simplified due to the unavailability of high-performance computing (HPC) resources.  

However, with engineering designs now becoming larger, more detailed and complex, the need for readily available HPC resources for the engineering modelling community has become increasingly evident. This is why the scalability of STAR-CCM+ is so significant.

For more information, on STAR-CCM+’s HPC scalability read: Simulation Sees Linear Processing Scalability to 55,000 Cores.