NVIDIA AI Operations - NCP-AIO FREE EXAM DUMPS QUESTIONS & ANSWERS]

Question 1

You are deploying an AI workload on a Kubernetes cluster that requires access to GPUs for training deep learning models. However, the pods are not able to detect the GPUs on the nodes.
What would be the first step to troubleshoot this issue?

A. Verify that the NVIDIA GPU Operator is installed and running on the cluster. B. Ensure that all pods are using the latest version of TensorFlow or PyTorch. C. Increase the number of CPU cores allocated to each pod to ensure better resource utilization. D. Check if the nodes have sufficient memory allocated for AI workloads.

Discussion 0

Correct Answer: A Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 2

You need to do maintenance on a node. What should you do first?

A. Set the node state to down in Slurm before completing maintenance. B. Disable job scheduling on all compute nodes in Slurm before completing maintenance. C. Set the node state to down in Slurm before completing maintenance. D. Drain the compute node using scontrol update.

Discussion 0

Correct Answer: D Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 3

A system administrator is troubleshooting a Docker container that crashes unexpectedly due to a segmentation fault. They want to generate and analyze core dumps to identify the root cause of the crash.
Why would generating core dumps be a critical step in troubleshooting this issue?

A. Core dumps capture the memory state of the process at the time of the crash. B. Core dumps provide real-time logs that can be used to monitor ongoing application performance. C. Core dumps restore the process to its previous state, often fixing the error-causing crash. D. Core dumps prevent future crashes by stopping any further execution of the faulty process.

Discussion 0

Correct Answer: A Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 4

A system administrator notices that jobs are failing intermittently on Base Command Manager due to incorrect GPU configurations in Slurm. The administrator needs to ensure that jobs utilize GPUs correctly.
How should they troubleshoot this issue?

A. Increase the number of GPUs requested in the job script to avoid using unconfigured GPUs. B. Check if MIG (Multi-Instance GPU) mode has been enabled incorrectly and reconfigure Slurm accordingly. C. Ensure that GPU resource limits have been correctly defined in Slurm's configuration file for each job type. D. Verify that non-MIG GPUs are automatically configured in Slurm when detected, and adjust configurations if needed.

Discussion 0

Correct Answer: B Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 5

A system administrator is looking to set up virtual machines in an HGX environment with NVIDIA Fabric Manager.
What three (3) tasks will Fabric Manager accomplish? (Choose three.)

A. Configures routing among NVSwitch ports. B. Coordinates with the GPU driver to initialize and train NVSwitch to GPU NVLink interconnects. C. Coordinates with the NVSwitch driver to train NVSwitch to NVSwitch NVLink interconnects. D. Installs GPU operator E. Installs vGPU driver as part of the Fabric Manager Package.

Discussion 0

Correct Answer: A,B,C Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 6

You are an administrator managing a large-scale Kubernetes-based GPU cluster using Run:AI.
To automate repetitive administrative tasks and efficiently manage resources across multiple nodes, which of the following is essential when using the Run:AI Administrator CLI for environments where automation or scripting is required?

A. Install the CLI on Windows machines to take advantage of its scripting capabilities. B. Use the CLI to manually allocate specific GPUs to individual jobs for better resource management. C. Ensure that the Kubernetes configuration file is set up with cluster administrative rights before using the CLI. D. Use the runai-adm command to directly update Kubernetes nodes without requiring kubectl.

Discussion 0

Correct Answer: C Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 7

You have successfully pulled a TensorFlow container from NGC and now need to run it on your stand- alone GPU-enabled server.
Which command should you use to ensure that the container has access to all available GPUs?

A. docker run nvcr.io/nvidia/tensorflow:<tag> B. docker run --gpus all nvcr.io/nvidia/tensorflow:<tag> C. kubectl create pod --gpu=all nvcr.io/nvidia/tensorflow:<tag> D. docker start nvcr.io/nvidia/tensorflow:<tag>

Discussion 0

Correct Answer: B Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).

Question 8

You are managing a Kubernetes cluster running AI training jobs using TensorFlow. The jobs require access to multiple GPUs across different nodes, but inter-node communication seems slow, impacting performance.
What is a potential networking configuration you would implement to optimize inter-node communication for distributed training?

A. Configure a dedicated storage network to handle data transfer between nodes during training. B. Use InfiniBand networking between nodes to reduce latency and increase throughput for distributed training jobs. C. Increase the number of replicas for each job to reduce the load on individual nodes. D. Use standard Ethernet networking with jumbo frames enabled to reduce packet overhead during communication.

Discussion 0

Correct Answer: B Vote an answer

Question 9

You are managing an on-premises cluster using NVIDIA Base Command Manager (BCM) and need to extend your computational resources into AWS when your local infrastructure reaches peak capacity.
What is the most effective way to configure cloudbursting in this scenario?

A. Set up a standby deployment in AWS and manually switch workloads to the cloud during peak times. B. Use BCM's Cluster Extension feature to automatically provision AWS resources when local resources are exhausted. C. Use BCM's built-in load balancer to distribute workloads evenly between on-premises and cloud resources without any pre-configuration. D. Manually provision additional cloud nodes in AWS when the on-premises cluster reaches its limit.

Discussion 0

Correct Answer: B Vote an answer

Explanation: Only visible for FreeCram members. You can sign-up / login (it's free).