최신 NCP-AIO 무료덤프 - NVIDIA AI Operations

문제1

You are tasked with deploying a deep learning framework container from NVIDIA NGC on a stand-alone GPU-enabled server.
What must you complete before pulling the container? (Choose two.)

A. Install Docker and the NVIDIA Container Toolkit on the server.

B. Set up a Kubernetes cluster to manage the container.

C. Generate an NGC API key and log in to the NGC container registry using docker login.

D. Install TensorFlow or PyTorch manually on the server before pulling the container.

정답: A,C

설명: (DumpTOP 회원만 볼 수 있음)

문제2

Your Kubernetes cluster is running a mixture of AI training and inference workloads. You want to ensure that inference services have higher priority over training jobs during peak resource usage times.
How would you configure Kubernetes to prioritize inference workloads?

A. Use Horizontal Pod Autoscaling (HPA) based on memory usage to scale up inference services during peak times.

B. Set up a separate namespace for inference services and limit resource usage in other namespaces.

C. Implement ResourceQuotas and PriorityClasses to assign higher priority and resource guarantees to inference workloads over training jobs.

D. Increase the number of replicas for inference services so they always have more resources than training jobs.

정답: C

설명: (DumpTOP 회원만 볼 수 있음)

문제3

A new researcher needs access to GPU resources but should not have permission to modify cluster settings or manage other users.
What role should you assign them in Run:ai?

A. L1 Researcher

B. Research Manager

C. Application Administrator

D. Department Administrator

정답: A

설명: (DumpTOP 회원만 볼 수 있음)

문제4

An administrator wants to check if the BlueMan service can access the DPU.
How can this be done?

A. Via a lightweight database operating in the DPU server

B. Via Linux dump files

C. Via system logs

D. Via the DOCA Telemetry Service (DTS)

정답: D

설명: (DumpTOP 회원만 볼 수 있음)

문제5

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. kubectl create pod --gpu=all nvcr.io/nvidia/tensorflow:<tag>

B. docker run --gpus all nvcr.io/nvidia/tensorflow:<tag>

C. docker run nvcr.io/nvidia/tensorflow:<tag>

D. docker start nvcr.io/nvidia/tensorflow:<tag>

정답: B

설명: (DumpTOP 회원만 볼 수 있음)

문제6

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. Use standard Ethernet networking with jumbo frames enabled to reduce packet overhead during communication.

B. Increase the number of replicas for each job to reduce the load on individual nodes.

C. Configure a dedicated storage network to handle data transfer between nodes during training.

D. Use InfiniBand networking between nodes to reduce latency and increase throughput for distributed training jobs.

정답: D

설명: (DumpTOP 회원만 볼 수 있음)

문제7

After completing the installation of a Kubernetes cluster on your NVIDIA DGX systems using BCM, how can you verify that all worker nodes are properly registered and ready?

A. Run kubectl get nodes to verify that all worker nodes show a status of "Ready".

B. Check each node manually by logging in via SSH and verifying system status with systemctl.

C. Run kubectl get pods to check if all worker pods are running as expected.

정답: A

설명: (DumpTOP 회원만 볼 수 있음)

최신 NCP-AIO 무료덤프 - NVIDIA AI Operations

우리와 연락하기

유용한 링크

최신 업데이트