NCP-ARI
Free NVIDIA-Certified Professional: AI Rack and Interconnect Practice Test
12 exam-style questions with full explanations — no sign-up. Score yourself, then close your gaps with the full course.
Exam fee ~$4007 exam domainsLevel Advanced2 timed practice tests in the course
✅ Free practice — no sign-up📝 Real exam-style questions💡 Detailed explanations💸 30-day money-back via Udemy
A technician is preparing a new GB200 NVL72 rack for installation. The rack's compute trays connect to a busbar that distributes 48V to each tray, and this busbar is fed by rectifier/power shelves that convert incoming facility power. How should this rack be classified for power planning purposes?
GB200 NVL72 racks use a DC power architecture: facility AC is converted by power shelves into 48V DC, which is then distributed to compute and switch trays via a busbar — this is the defining characteristic of a DC rack versus a traditional AC rack that uses per-outlet rPDUs.
A data center is preparing to receive a GB200 NVL72 rack, which weighs approximately 1,360 kg (3,000 lbs) fully populated. The facilities team confirms the raised floor tiles are rated for 600 kg/m² static load. Before scheduling delivery, what is the FIRST site-readiness step the deployment team should perform?
Structural floor load verification is a foundational physical-readiness check that must be confirmed before a multi-ton rack can be safely rolled in and positioned; failing this check can halt or damage the deployment before any electrical or network work begins.
A technician is preparing an extra-wide GB200 NVL72 rack for installation and needs to confirm the front-post-to-front-door clearance before mounting compute trays. Which clearance value should the technician verify meets NVIDIA's rack preparation requirement for full door swing and hot-swap tray access?
NVIDIA's rack preparation guidance for GB200 NVL72 extra-wide racks specifies approximately 1200 mm (47 in) of front clearance so the cabinet door can fully open and compute/switch trays can be serviced or hot-swapped without obstruction.
A high-density AI cluster requires NVIDIA NDR InfiniBand connectivity between two rows of racks located 1.5 km apart on the same data-center campus. Which transceiver and fiber combination BEST meets this requirement?
400G-FR4 optics are rated for up to 2 km over single-mode fiber, comfortably covering the 1.5 km run.
A technician is planning the overhead cable path for a new row of GB200 NVL72 racks, each requiring hundreds of copper InfiniBand DAC cables plus power whips running horizontally above the row. The bundled weight per linear foot is significantly higher than a typical data hall row. Which approach BEST addresses this load before installation begins?
Ultra-high-density AI clusters require an engineered support system sized to the actual calculated cable load, not an ad hoc solution. Trapeze/ladder racks are structurally rated for heavy bundled loads and their support spacing must be validated against the calculated weight per section.
A newly installed InfiniBand fiber link shows a link-up status but is experiencing intermittent packet retransmissions with a bit error rate (BER) above the acceptable threshold, even though the connector appears properly seated. Which troubleshooting step should be performed FIRST?
Contaminated or dirty connector end-faces are the most common and least invasive cause of link-up-but-high-BER symptoms. Inspection and cleaning is the standard first troubleshooting step before more drastic hardware changes.
A technician is assigned to service an rPDU feeding a live GB200 NVL72 rack drawing over 120 kW. What is the correct first step before beginning any work on the unit?
Standard electrical safety practice requires a formal lockout/tagout (LOTO) procedure: isolate the energy source, apply lock and tag, and confirm zero energy state with a rated tester before any hands-on work begins. This is mandatory in high-power (120 kW+) AI rack environments where residual energy or misidentified circuits pose lethal risk.
An AC-powered AI rack uses three-phase intelligent rPDUs. Monitoring shows Phase L1 is drawing significantly more current than L2 and L3, though total rack power is within capacity. What is the correct corrective action?
Intelligent rPDUs report per-phase current draw so technicians can redistribute equipment power cords across L1, L2, and L3, preventing any single phase from being overloaded while total capacity remains sufficient.
During a pre-deployment site survey for a row of liquid-cooled AI racks, the assessment team measures 900mm of clearance in the cold aisle and 1200mm in the hot aisle. Per standard high-density rack deployment practices for service access and airflow/CDU maintenance, what should the survey document as a finding?
Site surveys must verify rack spacing meets minimum clearance requirements for technician access, cable management, and CDU/manifold servicing; inadequate cold aisle clearance is a common deficiency that must be identified and remediated before installation to avoid rework.
A site team is routing AC power whips and InfiniBand fiber trunks into a GB200 NVL72 rack row. To mitigate EMI-induced signal degradation on the high-speed fiber links, what cable routing practice should be followed during rack preparation?
Rack preparation guidance requires dedicated, separated pathways (e.g., vertical duct for power on one side, horizontal duct for data on the other) with a specified minimum lateral separation to reduce electromagnetic interference that can degrade high-speed InfiniBand/fiber signal integrity.
Which transceiver form factor is used on NVIDIA Quantum-2 NDR InfiniBand switch ports to support the higher thermal and power requirements of 400/800Gb optical modules?
OSFP has a larger form factor with better thermal dissipation, allowing it to support the higher-power optics required for NDR/XDR speeds.
Which overhead cable support structure is specifically designed to distribute the weight of large bundled cable runs across multiple support points in a high-density AI cluster row?
Trapeze/ladder racks use rung-style construction supported at regular structural intervals, giving them the load rating needed to carry heavy bundled AI cluster cabling (power whips plus dense InfiniBand/AOC bundles) safely across a row.
Ready for the real thing?
The full course has two full-length practice tests, video lessons for every exam domain, hands-on labs and detailed answer explanations.
Start my full course on Udemy →NCP-ARI exam — quick answers
How much does the NCP-ARI exam cost?
The exam fee is approximately $400 and varies by region — confirm current pricing with the certification vendor before you book.
What topics are on the exam?
It covers 7 domains: AI Infrastructure Basics (20%), Pre-Deployment Planning and Site Assessment (11%), Rack Infrastructure Preparation (10%), High-Density Cabling Installation (30%), Cable Support Systems and Weight Management (7%), Testing, Verification, and Documentation (12%), Safety, Standards, and Compliance (10%). The full course has a dedicated chapter, lab and practice-test coverage for each.
Is this practice test really free?
Yes — all questions on this page are free with explanations and no sign-up. The paid Udemy course adds two full-length timed exams, video lessons and hands-on labs.
Will this prepare me for the real exam?
The questions mirror the real exam's style and are mapped to the official domains. This is exam-focused preparation — combine the free test with the full course's timed simulations to gauge your readiness.