NVIDIA AI Infrastructure and Operations Operations and Monitoring NCA-AIIO
Quiz 7 Question 3 of 20

In an enterprise AI training cluster, GPUs often sit idle, waiting for training data to be read from storage and loaded into memory—a state known as I/O starvation. To maximize GPU utilization and data throughput, which storage hardware strategy should you implement?

Select an answer to reveal the explanation.