Keywords: StarkStrom, GPU Infrastructure, Machine Learning, Simulation, Autonomous Driving, Aerodynamic, YOLO, Workstation, University, Engineering, Competition.
The collaboration between Genesis Cloud and StarkStrom Augsburg e.V. is a compelling example of how cutting-edge GPU cloud infrastructure can drive innovation and boost performance in competitive engineering projects. StarkStrom, a student-led initiative at Augsburg University of Applied Sciences in Germany, focuses on designing and building high-performance e-Formula racing cars.
In this interview with Felix Golatofski, CEO & Infrastructure Manager at StarkStrom, we explore how Genesis Cloud has enabled the project to overcome challenges, enhance productivity, and stay competitive in Formula Student competitions.
StarkStrom Augsburg. is the leading interdisciplinary project at the Augsburg University of Applied Sciences focused since 2011 on designing and building race cars to compete in the international Formula Student competitions. In this competitive environment, we participate in dynamic disciplines that test acceleration, performance, and efficiency, as well as static disciplines that assess design decisions, negotiation skills, and business planning. The unique blend of corporate-like structures and team spirit fosters high engagement and significant personal development among our team members, turning them into highly sought-after professionals. Our team, consisting of 40 to 70 students, participates in both driver-operated and autonomous vehicle categories.
Critical to our success is the use of FEM and FEA simulations, as well as machine learning networks, to optimize the performance of our e-Formula cars. These simulations are particularly essential for the aerodynamic design and autonomous driving technology. In addition, our autonomous driving technology has progressed significantly and we are proud to have entered the top 10 in the autonomous disciplines. This year we are focusing on further developing our YOLO cone detection and perception network, which requires state-of-the-art computing resources to maintain our competitive edge.
Before integrating Genesis Cloud's GPU infrastructure, we faced significant limitations with our local resources. Our powerful workstations were often occupied, which restricted our ability to run complex, resource-intensive simulations. We needed a solution that allowed us to conduct complex simulations iteratively, and with rapid turnaround times, which is critical for our project's success.
Our primary requirements were high RAM capacity and numerous CPU cores for aerodynamic simulations, along with powerful GPUs to accelerate these computations. Given the high complexity of our simulations, quick turnaround times were essential.
Integrating Genesis Cloud’s GPU infrastructure into our workflow has been transformative. Team members can now offload large and resource-intensive simulations to Genesis Cloud, which has freed up our local resources and streamlined our workflow. We provision Genesis Cloud instances as needed and equip them with the necessary software tools and configurations using an Ansible Playbook. This setup has allowed us to run Siemens StarCCM+ or Ansys Fluids simulations seamlessly. Once the simulations are complete, team members can easily power down the instances, optimizing our resource usage. This process ensured that our team members could quickly and efficiently access the computational power they needed, without interrupting the workflow of others.
Since adopting Genesis Cloud, we’ve seen significant gains in both efficiency and performance. Previously, some of our simulations would consume too much RAM and fail to produce results. With Genesis Cloud, our simulations now run to completion, and the quick simulation times have been crucial in high-pressure situations where results are needed fast. The ability to complete simulations in a fraction of the time has allowed us to stay on schedule and meet critical milestones.
Genesis Cloud’s GPU infrastructure has been instrumental in driving innovation in our autonomous vehicle technology. We are currently upgrading from YOLOv5 to YOLOv9 for our cone detection and perception systems, which demands extensive testing and a powerful system. Genesis Cloud has provided us with the computational power necessary to carry out these developments, helping us stay at the forefront of autonomous driving technology in the Formula Student competition.
Looking ahead, we plan to further integrate Genesis Cloud's services into our workflow by offloading more simulations to their infrastructure. We also aim to integrate these simulation tools with our local workstations so that workloads can be directly submitted to Genesis Cloud’s powerful computing environment. Our upcoming projects will continue to focus on the development of our aerodynamic package and other critical components of the car, with Genesis Cloud playing a key role in enabling the simulations and data collection necessary for presenting a competitive vehicle at Formula Student events.
Our experience with Genesis Cloud has been overwhelmingly positive. The Genesis Cloud team has been responsive and helpful, addressing our integration questions promptly and ensuring a smooth experience. The resource provisioning and user interface are easy to use, and we are excited about the upcoming plan to test provisioning via Terraform and Ansible, which will further enhance our workflow integration. We look forward to continuing our collaboration with Genesis Cloud and pushing the boundaries of what we can achieve in Formula Student.
Through its collaboration with Genesis Cloud, StarkStrom Augsburg has been able to overcome significant computational challenges, enhance its workflow, and push the boundaries of innovation in autonomous driving e-Formula racing. With Genesis Cloud’s GPU infrastructure, StarkStrom continues to thrive in the highly competitive sports environment, setting new benchmarks for university engineering projects.
The Genesis Cloud team
Never miss out again on Genesis Cloud news and our special deals: follow us on Twitter, LinkedIn, or Reddit.
Sign up for an account with Genesis Cloud here. If you want to find out more, please write to contact@genesiscloud.com.
