Posted On: Oct 14, 2021

AWS RoboMaker, a service that allows customers to simulate robotics applications at cloud scale, now supports GPU based simulation jobs for compute-intensive simulation workloads such as high fidelity simulation, vision processing, and machine learning (ML). Previously, AWS RoboMaker simulation jobs ran only on central processing unit (CPU) instances; now you can choose between a CPU based or GPU based simulation job.

Using AWS RoboMaker, developers can run, scale, and automate GPU based simulations. GPU based simulations support higher frames-per-second, higher resolutions, lower sensor latencies, and faster simulation job completion times than CPU based simulation jobs. These capabilities enable improved sensing by cameras and realistic rendering needed for use cases such as ML model training, reinforcement learning, and testing use cases that require high fidelity simulations. When running a GPU based simulation job, the AWS RoboMaker GUI tool viewer now supports higher resolutions, enabling you to see simulated objects in greater detail.

AWS RoboMaker is available in US East (N. Virginia), US East (Ohio), US West (Oregon), EU (Ireland), EU (Frankfurt), Asia Pacific (Tokyo), and Asia Pacific (Singapore) regions. Learn more about AWS RoboMaker GPU pricing on our webpage or get started by running a sample GPU application in the AWS RoboMaker console.