Turbonomic Optimizes GPU Workloads on Amazon EKS and ROSA
IBM Turbonomic can optimize GPU workloads on Amazon EKS and ROSA to improve GPU utilization and reduce costs. When inference workloads don't fully use their allocated GPU capacity, it results in wasted spend.
The company delivers the compute power needed for these workloads with GPU instances in the P5en, P5, and G6e families of Amazon EC2.
IBM Turbonomic continuously analyzes application demand against GPU supply in clusters and generates optimization actions such as scaling, resizing, and rebalancing to maintain performance while improving efficiency.