For computer-aided engineering, Google Cloud HPC can speed up your design and simulation processes
Mechanical Engineering improve creative groups are coming under ever greater stress to come up with solutions quickly, optimize the performance of goods, effectively shorten the time it takes to in the fiercely competitive marketplace of today.
Numerous CAE operations need a large amount of processing power. Google Cloud HPC is used by organizations to manage big datasets and complicated simulations that were previously handled by specialized, on-premises HPC facilities. But the cloud’s capacity to handle HPC workloads has significantly improved, enabling engineering teams to take use of the cloud’s flexibility, scalability, and performance.
Google has created a Computer Aided Engineering system that combines the necessary technology to effectively operate large CAE applications. The solution is designed to fit the simulation and analysis phases of CAE workflows and makes use of Google Cloud HPC capabilities.
Using CAE analysis and simulation to unleash creativity
Through a capacity to visually model, simulate, examine, and enhance designs, computer-aided engineering has entirely transformed the design and construction process, replacing the demand for physical prototypes and speeding up product development cycles. A vast array of applications, each tackling distinct technical issues, are covered by CAE.
These applications include:
Through the use of fluid dynamics, engineers may analyze heat transfer characteristics and optimize aerodynamic performance by simulating fluid flow around objects.
Thermal analysis ensures thermal efficiency and guards against overheating by simulating heat transport inside and around components.
Designing antennas, RF circuits, and electromagnetic compatibility all depend on the simulation of electromagnetic fields and their interactions with materials, which is made possible by electromagnetic analysis.
These use cases, along with a plethora of others, serve as the foundation of CAE, giving engineers strong tools to improve designs and guarantee the performance and safety of their products. CAE is becoming a crucial part of the engineering toolkit due to the complexity of engineering problems growing and the need for quick innovation.
Conceive: Using Computer Aided Design (CAD) software, engineers produce and investigate design ideas during the conceive stage.
Design: Engineers use CAD technologies to improve and streamline their designs throughout the design phase.
Develop: Using the techniques from the validate stage, engineers use CAE tools to build prototypes and test them in the develop stage.
Validate: During the validate phase, engineers confirm that their designs satisfy the necessary performance and safety requirements using CAE tools.
Manufacture: Using the CAD design as an input to multiple manufacturing processes, the manufacture step brings the designed and verified digital objects to life.
Google Cloud HPC is used extensively in the validate stage analysis and simulation. Engineers may run the validation stage more quickly or even run more faithful models when HPC is easily accessible for CAE processes, which boosts productivity and results in better products overall.
Google Cloud HPC: An effective way to speed up CAE processes
Google Cloud is assisting clients in setting up HPC systems that integrate the advantages of an elastic, flexible, planet-scale cloud with the HPC needed for CAE simulation and analysis.
Google have put together the appropriate cloud components to satisfy the demands of these computationally demanding workloads, making it simple to utilize Google Cloud for CAE processes. The H3 and C3 virtual machine families from Google Cloud, which are built on the newest Intel Xeon processors and provide balanced memory/flop ratios and excellent memory bandwidth, are the foundation of her solution architecture. These processors are ideal for CAE applications.
Up to 16GB of RAM per core may be used by the system to manage memory-intensive workloads and closely connected MPI applications. Additionally, it offers a range of storage options to meet both common and unusual I/O needs. It supports schedulers like Altair’s PBS professional and SchedMD’s Slurm for resource control.
It has been confirmed that the CAE Reference Architecture blueprint works well and is interoperable with the most popular CAE software, including Siemens Simcenter STAR-CCM+, Altair Radioss, Ansys Fluent, Ansys Mechanical, and Ansys LS-DYNA. The solution architecture is as follows:
Apart from the blueprint for the CAE reference architecture, Google also provide blueprints that show how to customize certain CAE software from source or binaries:
Siemens Star-CCM+ Ansys OpenFOAM Smooth Operation for demanding CAE workloads
The Google CAE solution is well suited for per-core-licensed applications since it makes use of Intel’s most recent generation Xeon processor, the Intel Sapphire Rapids, which is geared for great per-core performance.
Google examined the performance of Google’s H3 virtual machines (VMs) to the C2 generation of Intel-based VMs for many important CAE applications. Prominent CAE programs run 2.8x, 2.9x, and 2.6x quicker on H3 VMs than on C2 VMs: Ansys Fluent 2022 R2, Siemens Simcenter STAR-CCM+ 18.02.008, and Altair Radioss 2022.3.
By investing more computing resources in a single simulation, engineers may get design validation findings more quickly. The performance improvement of Ansys Fluent executing the 140 million cell F1 RaceCar CFD benchmark is shown in the graph below. The speedup doubles when the number of H3 VMs is increased from two to four. For Ansys Fluent, the comparable speedup is obtained with around 90% parallel efficiency even at 16 VMs, or 1408 cores.
Utilizing Cloud HPC Toolkit Blueprints to expedite CAE installations
In order to facilitate the easy integration of the CAE solution for bespoke deployments, google have released the general purpose CAE reference architecture as an Infrastructure as Code (IaC) blueprint for clients and system integrators. The CAE reference architecture may be easily deployed on Google cloud using this blueprint in conjunction with the open-source Cloud HPC Toolkit from Google.
Prominent CAE software, including Siemens Simcenter STAR-CCM+, Altair Radioss, Ansys Fluent, Ansys Mechanical, Ansys LS-DYNA, and OpenFOAM, has been used to evaluate the CAE solution.
Google provide comprehensive benchmark data, tools for getting started, and additional information on the solution’s components in her technical reference guide, Running computer-aided engineering workloads on Google Cloud.
For complex CAE processes, Google Cloud HPC offers a robust and scalable HPC solution. You may easily start this trip and discover the revolutionary potential of rapid simulations and analyses with her CAE solution.