McAfee-Secured Website

Exam Code: NCP-AII

Exam Name: NCP-AI Infrastructure

Certification Provider: NVIDIA

Questions & Answers

100% NVIDIA NCP-AII Guaranteed Success With Testking Training Materials

Get Real NCP-AII Exam Questions & Expert Verified Answers as Experienced on the Real Exam!

Questions & Answers

170 Questions & Answers with Testing Engine

"NCP-AI Infrastructure Exam", also known as NCP-AII exam, is a NVIDIA certification exam.

Pass your tests with the always up-to-date NCP-AII Exam Engine. Your NCP-AII training materials keep you at the head of the pack!

Satisfaction Guaranteed

Money Back Guarantee

Testking provides hassle-free money back guarantee with our products. That is because we have 100% trust in the abilities of our professional and experience product team, and our record is a proof of that.

99.6% PASS RATE
Was: $137.49
Now: $124.99

NVIDIA NCP-AII Practice Test 21

Question 200

During a GPU datacenter maintenance window, an engineer discovers: An AI cluster relies on BlueField offload for tenant isolation, but troubleshooting captures only host-side traffic and finds no explanation for dropped packets. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Gate DPU firmware rollout on signed images, mode compatibility, canary validation, and a tested recovery path.
  2. Document per-layer host, BlueField, and switch observability paths for any offloaded networking feature.
  3. Record DPU operating mode, owner, management interface, and supported configuration baseline at server commissioning.
  4. Require device rediscovery, allocation reconciliation, and job-drain checks whenever MIG partitions are reconfigured.

Explanation

Correct Answer

B:

This preventive practice addresses the underlying failure (Data-plane offload can move packet processing and filtering out of the host view, hiding the true drop location.) and supports continued assurance through: A controlled denied flow is attributable to the intended DPU enforcement rule and an allowed tenant flow continues as designed.

Incorrect Answers

A:

This prevention control targets another operational risk: A security team changes BlueField firmware on all AI servers but does not retain a recovery image, and one canary device fails to complete boot. The currently reported mechanism remains: Data-plane offload can move packet processing and filtering out of the host view, hiding the true drop location.

C:

This prevention control targets another operational risk: A GPU rack uses BlueField adapters with an unexpected embedded CPU management mode, and host traffic bypasses the data-plane policies administrators expected on the DPU. The currently reported mechanism remains: Data-plane offload can move packet processing and filtering out of the host view, hiding the true drop location.

D:

This prevention control targets another operational risk: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. The currently reported mechanism remains: Data-plane offload can move packet processing and filtering out of the host view, hiding the true drop location.

Question 201

A GPU cluster commissioning engineer investigates this situation: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. Which corrective deployment or operational action best addresses the stated problem?

  1. Drain affected workloads and use the documented enable/disable and instance creation sequence before returning capacity to service.
  2. Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits.
  3. Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change.
  4. Confirm the supported BlueField mode and deploy the intended host/DPU ownership and network acceleration configuration through approved utilities.

Explanation

Correct Answer

B:

For the stated NVIDIA AI infrastructure condition (Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU.), this action addresses the concrete failure: Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits. The underlying mechanism is: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator. Verification would require: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

Incorrect Answers

A:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The present issue is more specifically caused by: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. The present issue is more specifically caused by: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

D:

This option could be appropriate for a different NVIDIA infrastructure problem (2.1: Configure and manage a BlueField network platform): A GPU rack uses BlueField adapters with an unexpected embedded CPU management mode, and host traffic bypasses the data-plane policies administrators expected on the DPU. The present issue is more specifically caused by: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

Question 202

A GPU cluster commissioning engineer investigates this situation: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. Which measured result most directly proves that this specific problem is resolved?

  1. After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.
  2. A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.
  3. The approved DPU management interface responds and a supported DPU service test completes with host and embedded versions recorded.
  4. A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

Explanation

Correct Answer

D:

This is the measurable acceptance result for the described case: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement. It validates the corrective operation (Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. It does not directly establish the required result for the current case: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

B:

This observation would help verify a different incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. It does not directly establish the required result for the current case: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

C:

This observation would help verify a different incident: A cluster's BlueField-3 interfaces appear on the PCIe bus, but the DPU-side operating system has not been initialized and remote DOCA services fail. It does not directly establish the required result for the current case: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

Question 203

A GPU cluster commissioning engineer investigates this situation: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. Which root cause or infrastructure principle best explains the observed failure?

  1. MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.
  2. Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed.
  3. MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.
  4. A visible PCIe device is not proof that the embedded DPU software environment has booted or is configured.

Explanation

Correct Answer

C:

This technical mechanism explains the reported condition: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator. The observed situation is: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. A requirement-aligned correction is: Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits.

Incorrect Answers

A:

This underlying cause is associated with another GPU platform incident: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. The observed case instead concerns: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

B:

This underlying cause is associated with another GPU platform incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The observed case instead concerns: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

D:

This underlying cause is associated with another GPU platform incident: A cluster's BlueField-3 interfaces appear on the PCIe bus, but the DPU-side operating system has not been initialized and remote DOCA services fail. The observed case instead concerns: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

Question 204

A GPU cluster commissioning engineer investigates this situation: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Require device rediscovery, allocation reconciliation, and job-drain checks whenever MIG partitions are reconfigured.
  2. Enforce maintenance windows, scheduler cordoning, and before/after MIG inventory checks for every layout change.
  3. Include independent BlueField OS and service readiness tests in physical network platform acceptance.
  4. Validate supported MIG geometry and workload sizing before offering shared accelerator partitions to tenants.

Explanation

Correct Answer

D:

This preventive practice addresses the underlying failure (MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.) and supports continued assurance through: A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.

Incorrect Answers

A:

This prevention control targets another operational risk: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. The currently reported mechanism remains: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

B:

This prevention control targets another operational risk: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The currently reported mechanism remains: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

C:

This prevention control targets another operational risk: A cluster's BlueField-3 interfaces appear on the PCIe bus, but the DPU-side operating system has not been initialized and remote DOCA services fail. The currently reported mechanism remains: MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.

Question 205

During NVIDIA AI factory acceptance, a technician observes: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. Which corrective deployment or operational action best addresses the stated problem?

  1. Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change.
  2. Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits.
  3. Drain affected workloads and use the documented enable/disable and instance creation sequence before returning capacity to service.
  4. Provision the supported BlueField software image and verify DPU-side network, firmware, and management readiness independently of host PCIe discovery.

Explanation

Correct Answer

A:

For the stated NVIDIA AI infrastructure condition (A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition.), this action addresses the concrete failure: Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change. The underlying mechanism is: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions. Verification would require: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

Incorrect Answers

B:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. The present issue is more specifically caused by: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The present issue is more specifically caused by: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

D:

This option could be appropriate for a different NVIDIA infrastructure problem (2.1: Configure and manage a BlueField network platform): A cluster's BlueField-3 interfaces appear on the PCIe bus, but the DPU-side operating system has not been initialized and remote DOCA services fail. The present issue is more specifically caused by: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

Question 206

During NVIDIA AI factory acceptance, a technician observes: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. Which measured result most directly proves that this specific problem is resolved?

  1. A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.
  2. A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.
  3. The repaired DPU returns to the approved firmware and operating mode, and unaffected production nodes remain manageable throughout the recovery.
  4. After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

Explanation

Correct Answer

D:

This is the measurable acceptance result for the described case: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity. It validates the corrective operation (Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. It does not directly establish the required result for the current case: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

B:

This observation would help verify a different incident: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. It does not directly establish the required result for the current case: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

C:

This observation would help verify a different incident: A security team changes BlueField firmware on all AI servers but does not retain a recovery image, and one canary device fails to complete boot. It does not directly establish the required result for the current case: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

Question 207

During NVIDIA AI factory acceptance, a technician observes: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. Which root cause or infrastructure principle best explains the observed failure?

  1. MIG creates bounded GPU compute and memory instances rather than several unrestricted copies of the full physical accelerator.
  2. Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed.
  3. DPU firmware upgrades affect privileged embedded systems whose recovery and compatibility requirements differ from ordinary host software updates.
  4. MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

Explanation

Correct Answer

D:

This technical mechanism explains the reported condition: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions. The observed situation is: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. A requirement-aligned correction is: Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change.

Incorrect Answers

A:

This underlying cause is associated with another GPU platform incident: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. The observed case instead concerns: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

B:

This underlying cause is associated with another GPU platform incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The observed case instead concerns: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

C:

This underlying cause is associated with another GPU platform incident: A security team changes BlueField firmware on all AI servers but does not retain a recovery image, and one canary device fails to complete boot. The observed case instead concerns: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

Question 208

During NVIDIA AI factory acceptance, a technician observes: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. Which repeatable operational safeguard best prevents recurrence of this failure?

  1. Enforce maintenance windows, scheduler cordoning, and before/after MIG inventory checks for every layout change.
  2. Require device rediscovery, allocation reconciliation, and job-drain checks whenever MIG partitions are reconfigured.
  3. Validate supported MIG geometry and workload sizing before offering shared accelerator partitions to tenants.
  4. Document per-layer host, BlueField, and switch observability paths for any offloaded networking feature.

Explanation

Correct Answer

B:

This preventive practice addresses the underlying failure (MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.) and supports continued assurance through: After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.

Incorrect Answers

A:

This prevention control targets another operational risk: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. The currently reported mechanism remains: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

C:

This prevention control targets another operational risk: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. The currently reported mechanism remains: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

D:

This prevention control targets another operational risk: An AI cluster relies on BlueField offload for tenant isolation, but troubleshooting captures only host-side traffic and finds no explanation for dropped packets. The currently reported mechanism remains: MIG topology changes can invalidate ephemeral index-based device placement and monitoring assumptions.

Question 209

An infrastructure administrator receives the following incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. Which corrective deployment or operational action best addresses the stated problem?

  1. Bind scheduling and monitoring to stable MIG device identifiers or reliably refreshed discovery after every MIG layout change.
  2. Select supported MIG profiles that match requested compute and memory requirements and explicitly document partition resource limits.
  3. Stop the rollout and recover the canary using vendor-supported boot image and management procedures before proceeding in bounded groups.
  4. Drain affected workloads and use the documented enable/disable and instance creation sequence before returning capacity to service.

Explanation

Correct Answer

D:

For the stated NVIDIA AI infrastructure condition (A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads.), this action addresses the concrete failure: Drain affected workloads and use the documented enable/disable and instance creation sequence before returning capacity to service. The underlying mechanism is: Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed. Verification would require: A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.

Incorrect Answers

A:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. The present issue is more specifically caused by: Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed.

B:

This option could be appropriate for a different NVIDIA infrastructure problem (2.2: Configure Multi-Instance GPU (MIG) for AI and HPC): Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. The present issue is more specifically caused by: Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed.

C:

This option could be appropriate for a different NVIDIA infrastructure problem (2.1: Configure and manage a BlueField network platform): A security team changes BlueField firmware on all AI servers but does not retain a recovery image, and one canary device fails to complete boot. The present issue is more specifically caused by: Changing physical GPU partitioning can invalidate live device allocations even when the host driver remains installed.

Question 210

An infrastructure administrator receives the following incident: A team reconfigures MIG geometry while production GPU jobs are active, causing disruptive device recreation and lost workloads. Which measured result most directly proves that this specific problem is resolved?

  1. A device inventory and representative job test show the intended MIG instances, resource quotas, and successful workload placement.
  2. After a controlled partition reconfiguration, the intended job maps to the approved GPU and MIG UUID with expected memory capacity.
  3. A controlled test shows the traffic follows the documented DPU enforcement path and host management and datapath visibility agree on the mode.
  4. A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.

Explanation

Correct Answer

D:

This is the measurable acceptance result for the described case: A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset. It validates the corrective operation (Drain affected workloads and use the documented enable/disable and instance creation sequence before returning capacity to service.) rather than a related but different system check.

Incorrect Answers

A:

This observation would help verify a different incident: Researchers enable Multi-Instance GPU on a supported accelerator but assume two slices share the same capacity and hardware topology as one full GPU. It does not directly establish the required result for the current case: A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.

B:

This observation would help verify a different incident: A scheduler assigns a training container to a MIG instance by a temporary index, and reconfiguration leaves the container using the wrong partition. It does not directly establish the required result for the current case: A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.

C:

This observation would help verify a different incident: A GPU rack uses BlueField adapters with an unexpected embedded CPU management mode, and host traffic bypasses the data-plane policies administrators expected on the DPU. It does not directly establish the required result for the current case: A staged reconfiguration returns the target MIG profile set, and no protected job experiences an unplanned partition reset.

NCP-AI Infrastructure (NCP-AII) Practice Tests