QARGOS cuts simulation time by up to 90% with HPC on AWS

As QARGOS developed its born-electric cargo two-wheeler platform for commercial applications, it faced a demanding engineering challenge: designing a vehicle capable of carrying more cargo while maintaining stability, safety, and performance. Engineering simulations were critical for validating these requirements and informing design decisions, but QARGOS' existing compute environment could not support the scale and fidelity required. Some simulations took several days to complete, while more exhaustive workloads had to be put on hold entirely, limiting the number of iterations engineers could run.

These limitations became more critical as the vehicle progressed toward certification and production. Significant design changes after certification could require recertification, making it important to resolve issues and finalize key decisions beforehand. With timelines becoming increasingly critical, setting up an on-premises HPC environment would have required additional time, specialized HPC expertise, and substantial upfront investment.

Alok Das, co-founder of QARGOS, says, “We needed a scalable alternative that could provide the high-performance computing required to shorten simulation times, support more design iterations, and help engineers make critical decisions before certification.”

QARGOS began working with AWS to move its computationally intensive engineering simulations to the cloud. AWS HPC specialists worked with the company to understand its simulation requirements and recommended an architecture suited to its needs. “We relied on AWS to recommend the right HPC architecture and the best solution for our needs,” says Das.

QARGOS uses AWS ParallelCluster to configure and manage its HPC environment and runs simulations on Amazon Elastic Compute Cloud (Amazon EC2) Hpc7a instances and Amazon EC2 Hpc8a instances. Amazon FSx for Lustre provides high-performance shared storage for its simulation data.

With support from the AWS HPC team, QARGOS integrated its engineering software and workloads and got the HPC environment up and running within 4 hours. “AWS stepped in and played a fantastic role in terms of the integration,” says Das. “The team helped us get the system up and running much faster.”

Using AWS HPC, QARGOS has reduced computational time for key engineering simulations by up to 90 percent. Detailed battery simulations that previously took 68–92 hours can now be completed in approximately 10–15 hours. Through explicit dynamic simulation, QARGOS evaluated shock loading and validated the battery pack shock-test fixture before manufacturing, supporting successful physical testing and certification on the first attempt. This gives engineers results much sooner and allows more iterations within the development cycle.

“AWS enabled QARGOS engineers to assess battery shock-loading behaviour early in development and achieve first-time-right physical validation,” says Chaitanya Gandhi, senior manager at QARGOS.

For vehicle speed-breaker analysis, QARGOS used explicit dynamic simulation with real-world vehicle data collected through data acquisition (DAQ) systems to assess durability and determine a predictable factor of safety under speed breakers and other impact loads. Computational time for this analysis decreased by about 90 percent compared with its on-premises compute infrastructure. The company estimates that savings vary by simulation complexity and workload, with projected savings of up to $10,000 per simulation. “AWS makes faster simulations feasible, allowing us to use higher-order elements for greater accuracy and better correlation with physical test results,” says Hrishikesh Agashe, senior simulation engineer at QARGOS.

The faster turnaround is helping QARGOS make critical design decisions sooner as the vehicle progresses through certification. Engineers can spend less time waiting for simulations to complete and more time improving the vehicle, developing new ideas, and broadening their skills. “We could run multiple rounds of simulations in a shorter time frame, helping us make design decisions much faster,” says Das.

The HPC collaboration has also helped build trust between QARGOS and AWS, creating a foundation for further collaboration in IoT, data analytics, AI, and machine learning. As its vehicles generate increasing volumes of telemetry and real-world operating data, QARGOS plans to use AI for predictive maintenance, intelligent diagnostics, and continuous vehicle performance optimization. By analyzing data from multiple operating conditions, AI could also support improved durability predictions, energy efficiency, safety, and next-generation vehicle design. “AWS gives us a scalable architecture and a relationship that can grow with us as our journey continues,” says Das.