Exam NCA-AIIO Topic 2 Question 123 Discussion

Actual exam question for NVIDIA's NCA-AIIO exam
Question #: 123
Topic #: 2
In an AI infrastructure setup using NVIDIA GPUs across multiple nodes, you notice that the inter-node communication latency is higher than expected during distributed training. Which networking feature or protocol is most likely responsible for reducing latency in this scenario?

Suggested Answer: C Vote an answer

InfiniBand with RDMA (Remote Direct Memory Access) is the most effective networking feature for reducing inter-node communication latency in distributed training on NVIDIA GPUs. InfiniBand, paired with RDMA, enables direct memory access between nodes, bypassing CPU overhead and achieving ultra-low latency and high bandwidth (e.g., 200 Gb/s), critical for GPU-to-GPU data transfers via NVLink or NCCL.
Option A (NAT) manages addressing, not latency. Option B (TCP/IP over Ethernet) has higher overhead than InfiniBand. Option D (VLAN segmentation) aids isolation, not speed. NVIDIA's DGX and cluster documentation recommend InfiniBand for distributed AI workloads.

by Hilary at Aug 21, 2026, 01:55 AM

Limited Time Offer

15%

Off

Get Premium NCA-AIIO Questions as Interactive Self Test Engine or PDF

Comments

Chosen Answer:
This is a voting comment (?) , you can switch to a simple comment.
Switch to a voting comment New
Nick name: Submit Cancel
A voting comment increases the vote count for the chosen answer by one.

Upvoting a comment with a selected answer will also increase the vote count towards that answer by one. So if you see a comment that you already agree with, you can upvote it instead of posting a new comment.
0
0
0
10