cert
cert-1
cert-2

Pass Cisco 300-640 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
300-640 Exam - Verified By Experts
300-640 Premium File

300-640 Premium File

$89.99
$98.99
  • Premium File 60 Questions & Answers. Last Update: Nov 10, 2026

Whats Included:

  • Latest Questions
  • 100% Accurate Answers
  • Fast Exam Updates
 
$98.99
$89.99
accept 10 downloads in the last 7 days
block-screenshots
Cisco 300-640 Practice Test 10

Question 90

The acceptance board must decide whether an AI-infrastructure response to gpu interconnect topology awareness is fit for controlled operation. The fictional operating facts are: A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. A deployment recreates the required accelerator placement after node drain without starving protected services.
  2. Multi-GPU training design must match collective-communication demands to supported local GPU links, topology and NIC placement rather than count GPU cards alone.
  3. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.
  4. Collective benchmark efficiency increases without violating platform support or GPU memory constraints.
  5. A hybrid workload needs explicit identity, encrypted transport, consistent model and data versions, and ownership of cross-site failure modes.
  6. Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency.

Explanation

Correct Answer

B:

A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Multi-GPU training design must match collective-communication demands to supported local GPU links, topology and NIC placement rather than count GPU cards alone. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.

C:

A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes. Multi-GPU training design must match collective-communication demands to supported local GPU links, topology and NIC placement rather than count GPU cards alone.

D:

A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Collective benchmark efficiency increases without violating platform support or GPU memory constraints. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.

Incorrect Answers

A:

A deployment recreates the required accelerator placement after node drain without starving protected services. It would not independently demonstrate that the control works: A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Collective benchmark efficiency increases without violating platform support or GPU memory constraints. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.

E:

A hybrid workload needs explicit identity, encrypted transport, consistent model and data versions, and ownership of cross-site failure modes. It is not the rule that supports the proposed change and acceptance test: A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Multi-GPU training design must match collective-communication demands to supported local GPU links, topology and NIC placement rather than count GPU cards alone. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.

F:

Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency. It does not implement the governing rule for this approval: A recommendation-model experiment spans accelerators across two CPU sockets and reports irregular all-reduce stalls despite spare FLOPS. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes. Collective benchmark efficiency increases without violating platform support or GPU memory constraints.

Question 91

A data-center AI architecture committee considers container and accelerator scheduling in its AI Fundamentals and Applications review. At the affected site: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. The committee is deliberately not selecting a vendor product or applying a live change yet. It needs the foundational reasoning that makes the next design decision defensible. Which ONE engineering principle directly governs this situation? Choose ONE answer.

  1. Select block, file or object access by application consistency, sharing and throughput needs; storage capacity alone does not determine suitability.
  2. Reliable AI delivery separates data validation, training, evaluation, artifact registry, deployment, telemetry and rollback with reproducible version identifiers.
  3. A retrieval-augmented assistant should fetch authorized current passages and ground its responses in those passages rather than rely on model memory alone.
  4. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Explanation

Correct Answer

D:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Incorrect Answers

A:

Select block, file or object access by application consistency, sharing and throughput needs; storage capacity alone does not determine suitability. It does not govern the condition described: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

B:

Reliable AI delivery separates data validation, training, evaluation, artifact registry, deployment, telemetry and rollback with reproducible version identifiers. It does not govern the condition described: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

C:

A retrieval-augmented assistant should fetch authorized current passages and ground its responses in those passages rather than rely on model memory alone. It does not govern the condition described: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Question 92

An infrastructure diagnostician is investigating container and accelerator scheduling for the AI Fundamentals and Applications team. The reported incident reads: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Competing teams attribute the symptoms to different layers of the AI stack. Before accepting a hypothesis or touching production settings, the investigator must choose a discriminating set of measurements. Which ONE evidence package is MOST relevant? Choose ONE answer.

  1. Trace dataset hash, training run, evaluation results, model registry tag, deployed image and production drift signals.
  2. Capture read/write pattern, shared file semantics, metadata rate, namespace operations, IOPS and bandwidth measurements.
  3. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.
  4. Measure per-token or per-request p95 and p99 latency, queue wait, active batch size, concurrency and GPU memory headroom under load.

Explanation

Correct Answer

C:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Incorrect Answers

A:

Trace dataset hash, training run, evaluation results, model registry tag, deployed image and production drift signals. That check would not establish the decisive facts for this condition: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

B:

Capture read/write pattern, shared file semantics, metadata rate, namespace operations, IOPS and bandwidth measurements. That check would not establish the decisive facts for this condition: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

D:

Measure per-token or per-request p95 and p99 latency, queue wait, active batch size, concurrency and GPU memory headroom under load. That check would not establish the decisive facts for this condition: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Question 93

During a bounded AI-platform change-control review concerning container and accelerator scheduling, the site owner reports: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. The proposed work must have a narrow implementation scope, first be validated against the failure evidence and avoid unrelated system-wide tuning. Which ONE intervention should the engineers pilot before any wider rollout? Choose ONE answer.

  1. Pilot the narrowly scoped inference workflow with a named owner and benchmark it against a simple deterministic baseline.
  2. Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency.
  3. Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes.
  4. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Explanation

Correct Answer

D:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Incorrect Answers

A:

Pilot the narrowly scoped inference workflow with a named owner and benchmark it against a simple deterministic baseline. That change would leave the reported problem unresolved: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

B:

Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency. That change would leave the reported problem unresolved: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

C:

Place tightly coupled workers on topology-compatible GPU groups and validate communication scaling before expanding nodes. That change would leave the reported problem unresolved: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Question 94

Two independent technical reviewers examine a AI Fundamentals and Applications problem involving container and accelerator scheduling. Their shared incident log states: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. One reviewer requires measurements capable of separating the observed problem from neighboring failure modes; the other requires a bounded action that follows those measurements. Choose EXACTLY TWO statements that together supply those two parts.

  1. Compare the user's query with retrieved document IDs, access scopes, citation coverage and the freshness of indexed versions.
  2. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.
  3. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.
  4. Trace dataset hash, training run, evaluation results, model registry tag, deployed image and production drift signals.
  5. Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency.

Explanation

Correct Answer

B:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

C:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

Incorrect Answers

A:

Compare the user's query with retrieved document IDs, access scopes, citation coverage and the freshness of indexed versions. It would not provide the evidence needed for the paired diagnosis and follow-up: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

D:

Trace dataset hash, training run, evaluation results, model registry tag, deployed image and production drift signals. It would not provide the evidence needed for the paired diagnosis and follow-up: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

E:

Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency. It is not the scoped response supported by the accompanying evidence: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

Question 95

The acceptance board must decide whether an AI-infrastructure response to container and accelerator scheduling is fit for controlled operation. The fictional operating facts are: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. Model training modifies learned parameters using a controlled training pipeline; serving a model for inference should not silently retrain it on each user request.
  2. A deployment recreates the required accelerator placement after node drain without starving protected services.
  3. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.
  4. Data loaders sustain the target sample rate while checkpoints restore correctly under a simulated host loss.
  5. Deploy a compact validated inference artifact with signed updates, bounded local buffering and a defined offline fallback.
  6. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Explanation

Correct Answer

B:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. A deployment recreates the required accelerator placement after node drain without starving protected services. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

C:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

F:

A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Incorrect Answers

A:

Model training modifies learned parameters using a controlled training pipeline; serving a model for inference should not silently retrain it on each user request. It is not the rule that supports the proposed change and acceptance test: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

D:

Data loaders sustain the target sample rate while checkpoints restore correctly under a simulated host loss. It would not independently demonstrate that the control works: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. A deployment recreates the required accelerator placement after node drain without starving protected services. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

E:

Deploy a compact validated inference artifact with signed updates, bounded local buffering and a defined offline fallback. It does not implement the governing rule for this approval: A Kubernetes training pod remains Pending because the scheduler selects nodes without the advertised GPU resource. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

Question 96

A data-center AI architecture committee considers container and accelerator scheduling in its AI Fundamentals and Applications review. At the affected site: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. The committee is deliberately not selecting a vendor product or applying a live change yet. It needs the foundational reasoning that makes the next design decision defensible. Which ONE engineering principle directly governs this situation? Choose ONE answer.

  1. Reliable AI delivery separates data validation, training, evaluation, artifact registry, deployment, telemetry and rollback with reproducible version identifiers.
  2. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.
  3. Select block, file or object access by application consistency, sharing and throughput needs; storage capacity alone does not determine suitability.
  4. A hybrid workload needs explicit identity, encrypted transport, consistent model and data versions, and ownership of cross-site failure modes.

Explanation

Correct Answer

B:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Incorrect Answers

A:

Reliable AI delivery separates data validation, training, evaluation, artifact registry, deployment, telemetry and rollback with reproducible version identifiers. It does not govern the condition described: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

C:

Select block, file or object access by application consistency, sharing and throughput needs; storage capacity alone does not determine suitability. It does not govern the condition described: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

D:

A hybrid workload needs explicit identity, encrypted transport, consistent model and data versions, and ownership of cross-site failure modes. It does not govern the condition described: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Question 97

An infrastructure diagnostician is investigating container and accelerator scheduling for the AI Fundamentals and Applications team. The reported incident reads: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Competing teams attribute the symptoms to different layers of the AI stack. Before accepting a hypothesis or touching production settings, the investigator must choose a discriminating set of measurements. Which ONE evidence package is MOST relevant? Choose ONE answer.

  1. Review official supported components, approved release matrix, topology, GPU-network-storage bottlenecks and site facilities.
  2. Review false-positive and false-negative cost, data quality, stakeholder workflow, baseline metrics and feasible human escalation.
  3. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.
  4. Record on-device model footprint, local input rates, offline duration, update signatures and safe fallback triggers.

Explanation

Correct Answer

C:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Incorrect Answers

A:

Review official supported components, approved release matrix, topology, GPU-network-storage bottlenecks and site facilities. That check would not establish the decisive facts for this condition: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

B:

Review false-positive and false-negative cost, data quality, stakeholder workflow, baseline metrics and feasible human escalation. That check would not establish the decisive facts for this condition: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

D:

Record on-device model footprint, local input rates, offline duration, update signatures and safe fallback triggers. That check would not establish the decisive facts for this condition: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Question 98

During a bounded AI-platform change-control review concerning container and accelerator scheduling, the site owner reports: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. The proposed work must have a narrow implementation scope, first be validated against the failure evidence and avoid unrelated system-wide tuning. Which ONE intervention should the engineers pilot before any wider rollout? Choose ONE answer.

  1. Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency.
  2. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.
  3. Gate model promotion on evaluation and policy checks, retain lineage and provide a tested rollback to the preceding known-good version.
  4. Pilot the narrowly scoped inference workflow with a named owner and benchmark it against a simple deterministic baseline.

Explanation

Correct Answer

B:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.

Incorrect Answers

A:

Match the training and checkpoint workflow to supported file or block protocols and benchmark access under representative concurrency. That change would leave the reported problem unresolved: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

C:

Gate model promotion on evaluation and policy checks, retain lineage and provide a tested rollback to the preceding known-good version. That change would leave the reported problem unresolved: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

D:

Pilot the narrowly scoped inference workflow with a named owner and benchmark it against a simple deterministic baseline. That change would leave the reported problem unresolved: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Question 99

Two independent technical reviewers examine a AI Fundamentals and Applications problem involving container and accelerator scheduling. Their shared incident log states: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. One reviewer requires measurements capable of separating the observed problem from neighboring failure modes; the other requires a bounded action that follows those measurements. Choose EXACTLY TWO statements that together supply those two parts.

  1. Establish least-privilege hybrid links, versioned artifact replication and fail-closed handling for stale data or disconnected sites.
  2. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.
  3. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events.
  4. Review official supported components, approved release matrix, topology, GPU-network-storage bottlenecks and site facilities.
  5. Review false-positive and false-negative cost, data quality, stakeholder workflow, baseline metrics and feasible human escalation.

Explanation

Correct Answer

B:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

C:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

Incorrect Answers

A:

Establish least-privilege hybrid links, versioned artifact replication and fail-closed handling for stale data or disconnected sites. It is not the scoped response supported by the accompanying evidence: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

D:

Review official supported components, approved release matrix, topology, GPU-network-storage bottlenecks and site facilities. It would not provide the evidence needed for the paired diagnosis and follow-up: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

E:

Review false-positive and false-negative cost, data quality, stakeholder workflow, baseline metrics and feasible human escalation. It would not provide the evidence needed for the paired diagnosis and follow-up: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Inspect pod resource requests, device-plugin visibility, node labels, quota, volume readiness and scheduling events. A deployment recreates the required accelerator placement after node drain without starving protected services.

Question 100

The acceptance board must decide whether an AI-infrastructure response to container and accelerator scheduling is fit for controlled operation. The fictional operating facts are: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Signoff cannot rely on a good-looking utilization graph, a change ticket alone or an unmeasured promise. The response must include an appropriate governing engineering principle, a scoped intervention and a separate measurable acceptance criterion. Choose EXACTLY THREE statements that meet those roles.

  1. Gate model promotion on evaluation and policy checks, retain lineage and provide a tested rollback to the preceding known-good version.
  2. A deployment recreates the required accelerator placement after node drain without starving protected services.
  3. Data loaders sustain the target sample rate while checkpoints restore correctly under a simulated host loss.
  4. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.
  5. Model training modifies learned parameters using a controlled training pipeline; serving a model for inference should not silently retrain it on each user request.
  6. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

Explanation

Correct Answer

B:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. A deployment recreates the required accelerator placement after node drain without starving protected services. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

D:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent.

F:

A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Incorrect Answers

A:

Gate model promotion on evaluation and policy checks, retain lineage and provide a tested rollback to the preceding known-good version. It does not implement the governing rule for this approval: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior. A deployment recreates the required accelerator placement after node drain without starving protected services.

C:

Data loaders sustain the target sample rate while checkpoints restore correctly under a simulated host loss. It would not independently demonstrate that the control works: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. A deployment recreates the required accelerator placement after node drain without starving protected services. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

E:

Model training modifies learned parameters using a controlled training pipeline; serving a model for inference should not silently retrain it on each user request. It is not the rule that supports the proposed change and acceptance test: A model-serving namespace launches opportunistic background jobs and exhausts shared GPU memory during a customer peak. AI orchestration should request actual GPUs, memory and storage resources and respect scheduling isolation instead of treating all worker nodes as equivalent. Define hardware-aware scheduling constraints, namespaces and resource quotas with documented preemption and restart behavior.

Implementing Cisco Data Center AI Infrastructure (DCAI) (300-640) Practice Tests