Skip to content
NCP-AII
Cluster Test and Verification
hard
Question 1 of 66

In a control-plane rollout for GPU workloads, the team must diagnose the situation before taking the next documented action. Which response best addresses GPU-only validation gap?

ATreat a single-node result as proof that multi-node collectives work across the cluster.
BRun HPL with any process count because HPL automatically maps ranks to GPUs optimally.
CAdd NVLink, container RDMA, node-level NCCL, and multi-node NCCL checks because GPU visibility alone does not validate the cluster fabric.
DProceed with a generic Linux cluster procedure without confirming the NVIDIA-specific documented prerequisite.

Educational Content — CertQnA practice questions are written against official exam objectives, covering the same domains tested on the real exam. All content is original and independent — not actual exam questions, not affiliated with any certification vendor. Learn more about our content policy

Discussion

Be the first to share your understanding of this concept

⚠️ Discussion is for concept clarification only. Do not share or request actual exam questions or answers.

Sign in to join the discussion