[Q40-Q62] Get 100% Passing Success With True NCP-AII Exam! [Sep-2026]

NO.40 A customer has just completed the first boot of their DGX system and is prompted to create an administrative user. What is the correct approach for setting up this user to ensure secure BMC and GRUB access?

Create a unique, strong, lower-case username and password that will be used for both BMC and GRUB access, avoiding default or weak credentials.

Create separate usernames for BMC and GRUB to maximize flexibility.

Skip the creation of a new user and retain the default admin account for BMC and GRUB access.

Use “sysadmin” as the username and a simple password for ease of management.

NO.41 You are installing four NVIDIAAIOO GPUs into a server designed for AI training. The server motherboard has multiple PCIe Gen4 x16 slots. However, the server’s power supply unit (PSU) only has three 8-pin PCIe power connectors available. What is the BEST course of action to ensure all GPUs receive adequate power?

Use a PCIe power splitter cable on one of the 8-pin connectors to power two GPUs.

Install only three GPUs and leave the fourth unpowered.

Replace the existing PSU with a higher wattage PSU that has at least four 8-pin PCIe power connectors.

Connect the GPUs using the motherboard’s internal SATA power connectors.

Underclock the GPUs significantly to reduce their power consumption below the available PSU capacity.

NO.42 You are an infrastructure engineer tasked with validating a new AI training cluster before releasing it to users.
Your team wants to perform a NeMo burn-in to ensure both hardware and software are reliable and ready for production workloads. Which of the following actions are required as part of a proper NeMo burn-in process?
Pick the 2 correct responses below.

Download a pre-trained NeMo model and use it for a quick accuracy check on a user dataset, then consider the burn-in complete if results are reasonable.

Test inference using the NeMo API and approve the environment if the model outputs valid predictions.

Run the configured NeMo training job repeatedly or for an extended duration, monitoring for errors, stalls, or performance drops across all GPUs and nodes.

Configure a representative NeMo training or pretraining recipe and set up an executor to launch the job across intended nodes and GPUs.

NO.43 You are using NVIDIA Spectrum-X switches in your A1 infrastructure. You observe high latency between two GPU servers during a large distributed training job. After analyzing the switch telemetry, you suspect a suboptimal routing path is contributing to the problem. Which of the following methods offers the MOST granular control for influencing traffic flow within the Spectrum-X fabric to mitigate this?

Adjust the Equal-Cost Multi-Path (ECMP) hashing algorithm globally on all switches.

Configure QOS (Quality of Service) policies to prioritize traffic from the high-latency GPU servers.

Implement Adaptive Routing (AR) or Dynamic Load Balancing (DLB) features available on the Spectrum-X switches to dynamically adjust paths based on network conditions.

Manually configure static routes on the Spectrum-X switches to force traffic between the GPU servers along a specific path.

Disable IPv6 to simplify routing decisions.

NO.44 What command is needed to measure BER (Bit Error Rate)?

mlxconfig -d <device> q

ethtool -S <device>

mlxlink -d <device> -c -e

mstflint -d <device> q full

NO.45 You’ve successfully deployed BlueField OS to your SmartNlC. You need to verify that the Mellanox Ethernet driver (mlx5) is loaded and functioning correctly. What command would you use to confirm this?

NO.46 An administrator is configuring node categories in BCM for a DGX BasePOD cluster. They need to group all NVIDIA DGX H200 nodes under a dedicated category for GPU-accelerated workloads. Which approach aligns with NVIDIA’s recommended BCM practices?

Avoid categories and configure each DGX node individually via CLI.

Create a new “dgx-h200” category, assign all DGX H200 nodes to it.

Use the existing “dgxnodes” category without modification, as it is preconfigured for all DGX systems.

Assign nodes to the “login” category to simplify Slurm integration.

NO.47 A system administrator receives an alert about a potential hardware fault on an NVIDIA DGX A100. The GPU performance seems degraded, and the system fans are operating loudly. What step should be recommended to identify and troubleshoot the hardware fault?

Run a deep learning workload to stress test the GPUs and check whether the issue persists.

Check the NVIDIA System Management Interface (nvidia-smi) for GPU status and temperatures.

Power drain then restart the DGX and check if the performance degradation resolves.

Increase the fan speed to maximum and check whether the performance improves.

NO.48 You are using MIG (Multi-lnstance GPU) on an NVIDIAAI 00 GPU within a Kubernetes cluster. You want to configure a pod to use a specific MIG instance. How do you define the GPU resource request in the pod’s YAML definition?

Specify the full GPU ID (e.g., ‘nvidia.com/gpu: 1’) in the resource request.

Specify the MIG device profile name (e.g., ‘nvidia.com/mig-l g.7gb: 1 in the resource request.

Use a generic resource request like ‘nvidia.com/gpu: 0.125’ to allocate a portion of the GPIJ.

Use a custom resource definition (CRD) to define MIG instances and reference it in the resource request.

Using the annotation ‘nvidia.com/gpu.product’ and the value of the particular GPU device you desire.

NO.49 You are trying to install the NVIDIA Container Toolkit on a Linux distribution that is not officially supported in the NVIDIA documentation.
The standard installation instructions using ‘apt’ or “yum’ fail. What is the most appropriate approach to proceed with the installation?

Attempt to install the NVIDIA drivers and CUDA toolkit manually, bypassing the NVIDIA Container Toolkit altogether.

Identify a similar, supported Linux distribution and adapt the installation instructions for that distribution, carefully considering potential compatibility issues.

Download the source code for the NVIDIA Container Toolkit and compile it manually.

Create a Docker container with a supported distribution and run the application inside the container.

Contact NVIDIA support and request a custom installation package for your distribution.

NO.50 A financial services firm is deploying an AI model for fraud detection that requires rapid inference and data retrieval across multiple sites. Which feature should their storage system prioritize?

High capacity with moderate speed.

Tape backup systems.

Low-cost HDD solutions.

Multi-protocol data access with low latency.

NO.51 You are an infrastructure engineer tasked with validating a new AI training cluster before releasing it to users. Your team wants to perform a NeMo burn-in to ensure both hardware and software are reliable and ready for production workloads. Which of the following actions are required as part of a proper NeMo burn-in process? (Choose two.)

Run the configured NeMo training job repeatedly or for an extended duration, monitoring for errors, stalls, or performance drops across all GPUs and nodes.

Configure a representative NeMo training or pretraining recipe and set up an executor to launch the job across intended nodes and GPUs.

Test inference using the NeMo API and approve the environment if the model outputs valid predictions.

Download a pre-trained NeMo model and use it for a quick accuracy check on a user dataset, then consider the burn-in complete if results are reasonable.

NO.52 An engineer needs to verify NVLink isolation on a single node with 8 GPUs. Which NCCL test configuration stresses switch bisection bandwidth?

Use NCCL_TESTS_SPLIT=”DIV 8″ with point-to-point tests

Use all_reduce_perf -b 8 -e 16G -f 2 -g 8 with NCCL_TESTS_SPLIT=”AND 0x1″

Use reduce_scatter_perf -b 8 -e 16G -f 2 -g 4

Use all_reduce_perf -b 8 -e 16G -f 2 -g 8 without splits

NO.53 You are tasked with automating the BlueField OS deployment process across a large number of SmartNICs. Which of the following methods is MOST suitable for this task?

Manually flashing each SmartNIC using the ‘bfboot utility on a workstation.

Using a network boot (PXE) server to deploy the BlueField OS image over the network. This allows centralized management and scalability.

Creating a custom ISO image with the BlueField OS and booting each SmartNIC from a USB drive.

Utilizing the ‘dd’ command to directly copy the image to each SmartNIC’s flash memory.

Utilizing a custom-built python script to flash each individual card, controlled from a central server. This method supports parallel flashing.

NO.54 You are configuring an NVIDIAAIOO GPU in a server, and after installation and driver setup, lower than the GPU’s specified TDP. What are the possible reasons for this? nvidia-smi reports a power limit much

The GPU is faulty.

The power supply is not providing enough power.

The driver is not correctly installed.

The system BIOS is limiting the power to the PCIe slot.

The GPIJ is in a low-power mode due to inactivity.

NO.55 After running a 24-hour stress test on a DGX node, the administrator should verify which two key metrics to ensure system stability?

Average CPU usage > 80% and Docker container uptime.

No thermal throttling events and consistent GPU utilization > 95% throughout the test.

SSD write endurance and RAM capacity.

Total energy consumption and NVLink bandwidth.

NO.56 A leaf switch shows “FW Version Mismatch” alerts for transceivers after cluster expansion. Which tool validates transceiver firmware against expected versions?

ethtool

iblinkinfo

mlxconfig

flint

NO.57 You are evaluating the integration of NVIDIA BlueField DPUs into your data center’s storage architecture to optimize AI workloads. The storage solution chosen has incorporated BlueField DPUs to enhance performance and efficiency. Which of the following benefits directly results from this integration?

Elimination of latency issues in data processing tasks.

Enhanced I/O performance with NVMe storage access speeds.

Reduced CPU load by offloading data processing tasks to DPUs.

Unlimited scalability by adding more DPUs without architectural changes.

NO.58 You are standing up an NVIDIA DGX system for enterprise production. Stakeholder teams require system reliability, performance consistency under load, and proper escalation processes before release. A recent system in another cluster experienced intermittent GPU failures attributed to missed early-stage validation.
Which deployment and validation sequence best addresses production readiness and mitigates the risk of avoidable downtime or performance loss?

Install latest OS images and drivers, confirm OS and container functionality, invite users for a monitored production trial, and collect workload feedback to plan any further diagnostics or updates.

Complete hardware and cabling, power on the system, update firmware and drivers, run full hardware health checks and stress diagnostics using NVSM, verify all GPU and system sensor logs, and validate GPU accessibility.

Update network topology, assign static IPs and DNS entries, register the system with NVIDIA, then conduct basic OS-level checks and enable user access after login testing is successful.

Power on the system, install all AI frameworks, configure the CUDA and library stack, set up user environments, then plan stress tests and diagnostics as part of ongoing routine operations.

NO.59 You have a server equipped with multiple NVIDIA GPUs connected via NVLink. You want to monitor the NVLink bandwidth utilization in real-time. Which tool or method is the most appropriate and accurate for this?

Using ‘nvidia-smi’ with the ‘-display=nvlink’ option.

Parsing the output of *nvprof during a representative workload.

Utilizing DCGM (Data Center GPU Manager) with its NVLink monitoring capabilities.

Monitoring network interface traffic using ‘iftop’ or ‘tcpdump’ .

Using ‘gpustat’ .

NO.60 You are responsible for ensuring interoperability between AI applications deployed across a diverse IT landscape, including an on-premises data center equipped with NVIDIA GPUs and multiple cloud platforms from different vendors. These environments need to support complex AI workflows that involve large-scale data processing, real-time analytics, and machine learning model training. To maintain consistent performance and flexibility, which strategy should you prioritize?

Choose a vendor and standardize on one storage solution across the environments to simplify management to improve interoperability

Ensure that all environments use compatible storage protocols and APIs, such as NFS or S3, to facilitate data exchange and integration across platforms.

Implement a multi-cloud strategy that leverages native storage solutions in each cloud platform while using middleware to ensure interoperability and data consistency.

Focus on increasing network bandwidth between locations to reduce latency and improve data transfer speeds.

NO.61 During a maintenance window, a system administrator needs to verify the CUDA version installed on the NVIDIA DGX server to ensure compatibility with applications. Which command could be used in the maintenance script to check the installed CUDA version?

dpkg -l | grep cuda

cat /bin/cuda/cuda.txt

nvidia-smi –query-cuda-version

nvcc-version

NO.62 An AI server with 8 GPUs is experiencing random system crashes under heavy load. The system logs indicate potential memory errors, but standard memory tests (memtest86+) pass without any failures. The GPUs are passively cooled. What are the THREE most likely root causes of these crashes?

Incompatible NVIDIA driver version with the installed Linux kernel.

GPIJ memory errors that are not detectable by standard CPU-based memory tests.

Insufficient airflow within the server, leading to overheating of the GPUs and VRMs.

A faulty power supply unit (PSU) that is unable to provide stable power under peak load.

Network congestion causing intermittent data corruption during distributed training.