cert
cert-1
cert-2

Pass NVIDIA NCP-AII Exam in First Attempt Guaranteed!

Get 100% Latest Exam Questions, Accurate & Verified Answers to Pass the Actual Exam!
30 Days Free Updates, Instant Download!

cert-5
cert-6
NCP-AII Exam - Verified By Experts
NCP-AII Premium File

NCP-AII Premium File

$89.99
$98.99
  • Premium File 170 Questions & Answers. Last Update: Nov 06, 2026

Whats Included:

  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
 
$98.99
$89.99
accept 14 downloads in the last 7 days
block-screenshots
NVIDIA NCP-AII Practice Test 85

Question 840

During an NVIDIA AI factory acceptance test, engineers discover: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports. Which TWO complementary elements are needed: the correct technical action and the strongest independent diagnostic evidence? Select TWO answers.

  1. Use multi-flow cluster-scale throughput tests to measure effective bisection and hotspot oversubscription.
  2. Inspect local device topology, nccl-tests rank count, all-reduce bandwidth and link error metrics.
  3. Assert a nonblocking AI network because every endpoint NIC independently negotiates 400 Gb/s.
  4. Inspect per-tier utilization, congestion counters, full-rack NCCL measurements and ECMP distribution.
  5. Align MPI ranks, CPU NUMA affinity and GPU allocation with the tested host and fabric topology.

Explanation

Correct Answers

A:

This is an appropriate correction for AI fabric bisection and oversubscription test: Use multi-flow cluster-scale throughput tests to measure effective bisection and hotspot oversubscription. Observed scenario: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports. The governing mechanism is: Simultaneous east-west traffic stresses shared fabric resources beyond individual host line-rate capability.

D:

This is the direct operational evidence to investigate: Inspect per-tier utilization, congestion counters, full-rack NCCL measurements and ECMP distribution. Observed scenario: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports. The governing mechanism is: Simultaneous east-west traffic stresses shared fabric resources beyond individual host line-rate capability.

Incorrect Answers

B:

This alternative is relevant to Intra-node NCCL all-reduce test scope but does not directly resolve the current AI fabric bisection and oversubscription test issue: Inspect local device topology, nccl-tests rank count, all-reduce bandwidth and link error metrics. Observed scenario: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports.

C:

This tempting but unsafe shortcut is not a sufficient remediation for AI fabric bisection and oversubscription test: Assert a nonblocking AI network because every endpoint NIC independently negotiates 400 Gb/s. Observed scenario: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports. The governing mechanism is: Simultaneous east-west traffic stresses shared fabric resources beyond individual host line-rate capability. The proportionate action is: Use multi-flow cluster-scale throughput tests to measure effective bisection and hotspot oversubscription. A direct check is: Inspect per-tier utilization, congestion counters, full-rack NCCL measurements and ECMP distribution. The acceptance condition is: The fabric meets aggregate traffic expectations with no persistent oversubscription hotspot under representative concurrent jobs.

E:

This alternative is relevant to Multi-process HPL rank placement and CPU affinity but does not directly resolve the current AI fabric bisection and oversubscription test issue: Align MPI ranks, CPU NUMA affinity and GPU allocation with the tested host and fabric topology. Observed scenario: A fabric upgrade adds GPUs faster than switch uplinks and performance drops despite normal ports.

Question 841

A production GPU cluster review identifies this condition: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing. Which operational response is MOST appropriate to address the root issue? Select ONE best answer.

  1. Run a controlled representative GPU stress with monitoring of temperatures, memory and power-limit behavior.
  2. Check transceiver identity, firmware and supported FEC or link settings at both ends before stress testing.
  3. Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability.
  4. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.

Explanation

Correct Answer

C:

This is an appropriate correction for InfiniBand link state LID and physical activation: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

Incorrect Answers

A:

This alternative is relevant to GPU thermal power and memory single-node burn-in but does not directly resolve the current InfiniBand link state LID and physical activation issue: Run a controlled representative GPU stress with monitoring of temperatures, memory and power-limit behavior. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

B:

That option concerns Transceiver firmware and negotiated link test. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

D:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.

Question 842

In a professional AI infrastructure deployment, a technician observes: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing. Which underlying infrastructure principle BEST explains the issue? Select ONE best answer.

  1. A representative performance benchmark is meaningful only when the tested hardware and software baseline are documented.
  2. InfiniBand requires subnet management for address assignment and routing across the switched fabric.
  3. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  4. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Explanation

Correct Answer

D:

This is the governing infrastructure principle: An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

Incorrect Answers

A:

That option concerns Cluster-wide preflight hardware inventory. The acceptance condition is: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

B:

That option concerns InfiniBand subnet manager and route distribution. The acceptance condition is: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

C:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Question 843

The cluster operations team needs to evaluate this situation: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing. Which diagnostic evidence should be obtained FIRST to substantiate the likely fault or configuration gap? Select ONE best answer.

  1. Inspect subnet-manager state, LIDs, partition policy, discovered topology and routing information.
  2. Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs.
  3. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  4. Inspect ib_write_bw or ib_write_lat output, MTU, negotiated link speed and HCA or switch counters.

Explanation

Correct Answer

B:

This is the direct operational evidence to investigate: Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

Incorrect Answers

A:

This alternative is relevant to InfiniBand subnet manager and route distribution but does not directly resolve the current InfiniBand link state LID and physical activation issue: Inspect subnet-manager state, LIDs, partition policy, discovered topology and routing information. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

C:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

D:

That option concerns InfiniBand verbs microbenchmark bandwidth latency. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Question 844

At a multi-node AI infrastructure handoff, the team records: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing. After corrective work, which observable result BEST confirms that the requirement is met? Select ONE best answer.

  1. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  2. Link errors stabilize below approved thresholds and NCCL performance recovers under repeated load.
  3. New and existing ports become correctly addressed and end-to-end fabric paths work after manager failover.
  4. All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

Explanation

Correct Answer

D:

This is an observable verified result, not merely a claimed configuration change: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

Incorrect Answers

A:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

B:

This alternative is relevant to InfiniBand link error and recovery counters but does not directly resolve the current InfiniBand link state LID and physical activation issue: Link errors stabilize below approved thresholds and NCCL performance recovers under repeated load. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

C:

This alternative is relevant to InfiniBand subnet manager and route distribution but does not directly resolve the current InfiniBand link state LID and physical activation issue: New and existing ports become correctly addressed and end-to-end fabric paths work after manager failover. Observed scenario: An HCA is detected by Linux but the connected InfiniBand port remains Down or Initializing.

Question 845

During an NVIDIA AI factory acceptance test, engineers discover: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. Which operational response is MOST appropriate to address the root issue? Select ONE best answer.

  1. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  2. Correlate cumulative and rate-based InfiniBand link counters with workloads and replace defective physical paths as indicated.
  3. Configure a supported active subnet manager and validated standby behavior with consistent fabric policy.
  4. Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability.

Explanation

Correct Answer

D:

This is an appropriate correction for InfiniBand link state LID and physical activation: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Incorrect Answers

A:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

B:

This alternative is relevant to InfiniBand link error and recovery counters but does not directly resolve the current InfiniBand link state LID and physical activation issue: Correlate cumulative and rate-based InfiniBand link counters with workloads and replace defective physical paths as indicated. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

C:

This alternative is relevant to InfiniBand subnet manager and route distribution but does not directly resolve the current InfiniBand link state LID and physical activation issue: Configure a supported active subnet manager and validated standby behavior with consistent fabric policy. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Question 846

A production GPU cluster review identifies this condition: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. Which underlying infrastructure principle BEST explains the issue? Select ONE best answer.

  1. RDMA verbs benchmarks reveal transport behavior not fully measured by ordinary TCP connectivity tools.
  2. DCGM diagnostic suites exercise hardware and software health more deeply than passive status queries, sometimes consuming substantial resources.
  3. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  4. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Explanation

Correct Answer

D:

This is the governing infrastructure principle: An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Incorrect Answers

A:

That option concerns InfiniBand verbs microbenchmark bandwidth latency. The acceptance condition is: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

B:

That option concerns DCGM active single-node diagnostic suites. The acceptance condition is: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

C:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Question 847

In a professional AI infrastructure deployment, a technician observes: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. Which diagnostic evidence should be obtained FIRST to substantiate the likely fault or configuration gap? Select ONE best answer.

  1. Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs.
  2. Inspect deviceQuery or equivalent CUDA sample output, driver-runtime versions and kernel launch error codes.
  3. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  4. Inspect GPU temperature and power traces, clock throttle reasons, error counters and completed stress duration.

Explanation

Correct Answer

A:

This is the direct operational evidence to investigate: Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Incorrect Answers

B:

This alternative is relevant to CUDA device query and kernel smoke tests but does not directly resolve the current InfiniBand link state LID and physical activation issue: Inspect deviceQuery or equivalent CUDA sample output, driver-runtime versions and kernel launch error codes. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

C:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. The governing mechanism is: An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup. The proportionate action is: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability.

D:

This alternative is relevant to GPU thermal power and memory single-node burn-in but does not directly resolve the current InfiniBand link state LID and physical activation issue: Inspect GPU temperature and power traces, clock throttle reasons, error counters and completed stress duration. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Question 848

The cluster operations team needs to evaluate this situation: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. After corrective work, which observable result BEST confirms that the requirement is met? Select ONE best answer.

  1. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  2. Each intended GPU successfully initializes and executes a minimal repeatable CUDA workload under its supported stack.
  3. Every intended host and switch port reaches the correct physical peer with documented redundant paths.
  4. All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

Explanation

Correct Answer

D:

This is an observable verified result, not merely a claimed configuration change: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Incorrect Answers

A:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet. The governing mechanism is: An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup. The proportionate action is: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability. A direct check is: Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs. The acceptance condition is: All required HCA links are Active with valid fabric membership and can reach intended peer endpoints.

B:

This alternative is relevant to CUDA device query and kernel smoke tests but does not directly resolve the current InfiniBand link state LID and physical activation issue: Each intended GPU successfully initializes and executes a minimal repeatable CUDA workload under its supported stack. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

C:

This alternative is relevant to Cable topology mapping with LLDP and physical records but does not directly resolve the current InfiniBand link state LID and physical activation issue: Every intended host and switch port reaches the correct physical peer with documented redundant paths. Observed scenario: A node's port shows Active yet its assigned LID or fabric membership is inconsistent with the subnet.

Question 849

At a multi-node AI infrastructure handoff, the team records: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists. Which operational response is MOST appropriate to address the root issue? Select ONE best answer.

  1. Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability.
  2. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  3. Capture a reproducible preflight inventory of node hardware, software, connections and expected acceptance tests.
  4. Execute supported CUDA runtime initialization and minimal kernel tests on each scheduled accelerator.

Explanation

Correct Answer

A:

This is an appropriate correction for InfiniBand link state LID and physical activation: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

Incorrect Answers

B:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

C:

This alternative is relevant to Cluster-wide preflight hardware inventory but does not directly resolve the current InfiniBand link state LID and physical activation issue: Capture a reproducible preflight inventory of node hardware, software, connections and expected acceptance tests. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

D:

This alternative is relevant to CUDA device query and kernel smoke tests but does not directly resolve the current InfiniBand link state LID and physical activation issue: Execute supported CUDA runtime initialization and minimal kernel tests on each scheduled accelerator. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

Question 850

During an NVIDIA AI factory acceptance test, engineers discover: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists. Which TWO complementary elements are needed: the correct technical action and the strongest independent diagnostic evidence? Select TWO answers.

  1. Inspect node inventory, GPU and NIC UUIDs, firmware and driver versions, cabling records and test eligibility.
  2. Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs.
  3. Schedule supported active DCGM diagnostics for the relevant GPU class and investigate detailed failures.
  4. Equate an InfiniBand PCIe device in lspci with a usable active subnet connection.
  5. Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability.

Explanation

Correct Answers

B:

This is the direct operational evidence to investigate: Inspect ibstat or equivalent device output, port state, LID, switch interface and subnet-manager logs. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

E:

This is an appropriate correction for InfiniBand link state LID and physical activation: Confirm InfiniBand physical and logical port states, subnet assignment and end-to-end reachability. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists. The governing mechanism is: An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

Incorrect Answers

A:

This alternative is relevant to Cluster-wide preflight hardware inventory but does not directly resolve the current InfiniBand link state LID and physical activation issue: Inspect node inventory, GPU and NIC UUIDs, firmware and driver versions, cabling records and test eligibility. Observed scenario: A technician starts training when a leaf switch port is physically disconnected although the OS interface exists.

C:

That option concerns DCGM active single-node diagnostic suites. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

D:

This tempting but unsafe shortcut is not a sufficient remediation for InfiniBand link state LID and physical activation: Equate an InfiniBand PCIe device in lspci with a usable active subnet connection. A technician starts training when a leaf switch port is physically disconnected although the OS interface exists. An enumerated HCA does not imply a fabric-ready port; InfiniBand state and LID assignment reflect additional setup.

NCP-AI Infrastructure (NCP-AII) Practice Tests