Cisco 300-640 Practice Test 28
Question 270
The acceptance board must decide whether an AI-infrastructure response to rack power budget for accelerators is fit for controlled operation. The fictional operating facts are: An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. 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.
- Quarterly growth simulations remain within capacity headroom and all retained runs can be traced.
- Efficiency proposals must distinguish total facility power from IT energy and compare measured PUE under equivalent loads.
- The rack survives an authorized single-feed test without exceeding the remaining electrical capacity.
- Package or shard files appropriately and scale the supported metadata path while preserving data semantics.
- Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required.
- High-density GPU racks must fit measured branch-circuit, PDU and failover electrical budgets under realistic power transients.
Explanation
Correct Answer
C:An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. The rack survives an authorized single-feed test without exceeding the remaining electrical capacity. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required.
E:An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required. High-density GPU racks must fit measured branch-circuit, PDU and failover electrical budgets under realistic power transients.
F:An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. High-density GPU racks must fit measured branch-circuit, PDU and failover electrical budgets under realistic power transients. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required.
Incorrect Answers
A:Quarterly growth simulations remain within capacity headroom and all retained runs can be traced. It would not independently demonstrate that the control works: An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. The rack survives an authorized single-feed test without exceeding the remaining electrical capacity. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required.
B:Efficiency proposals must distinguish total facility power from IT energy and compare measured PUE under equivalent loads. It is not the rule that supports the proposed change and acceptance test: An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. High-density GPU racks must fit measured branch-circuit, PDU and failover electrical budgets under realistic power transients. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required.
D:Package or shard files appropriately and scale the supported metadata path while preserving data semantics. It does not implement the governing rule for this approval: An AI colocation pod has dual PSUs but both are connected to the same overloaded feed. Right-size and distribute accelerator nodes across validated A/B feeds with capped power policy where required. The rack survives an authorized single-feed test without exceeding the remaining electrical capacity.
Question 271
A data-center AI architecture committee considers cooling and thermal headroom in its AI Infrastructure Components and Architecture review. At the affected site: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. 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.
- AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
- Training storage must sustain aggregate sequential and random reads across workers, not just advertised peak bandwidth.
- Distributed AI placement must limit correlated failures and balance locality against rack, host and power dependencies.
- Training, management, tenant and sensitive data paths need enforceable policy boundaries independent of aggregate throughput.
Explanation
Correct Answer
A:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Incorrect Answers
B:Training storage must sustain aggregate sequential and random reads across workers, not just advertised peak bandwidth. It does not govern the condition described: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
C:Distributed AI placement must limit correlated failures and balance locality against rack, host and power dependencies. It does not govern the condition described: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
D:Training, management, tenant and sensitive data paths need enforceable policy boundaries independent of aggregate throughput. It does not govern the condition described: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Question 272
An infrastructure diagnostician is investigating cooling and thermal headroom for the AI Infrastructure Components and Architecture team. The reported incident reads: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. 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.
- Record adapter firmware, host kernel, accelerator driver, CUDA/runtime level, Cisco UCS profiles and network release compatibility.
- Count metadata operations, open latency, namespace contention and worker-level file listing times.
- Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
- Observe GPU idle gaps, CPU saturation, page faults, dataloader queue depth, NUMA affinity and PCIe locality.
Explanation
Correct Answer
C:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Incorrect Answers
A:Record adapter firmware, host kernel, accelerator driver, CUDA/runtime level, Cisco UCS profiles and network release compatibility. That check would not establish the decisive facts for this condition: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
B:Count metadata operations, open latency, namespace contention and worker-level file listing times. That check would not establish the decisive facts for this condition: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
D:Observe GPU idle gaps, CPU saturation, page faults, dataloader queue depth, NUMA affinity and PCIe locality. That check would not establish the decisive facts for this condition: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Question 273
During a bounded AI-platform change-control review concerning cooling and thermal headroom, the site owner reports: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. 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.
- Use an approved durable checkpoint repository with atomic or validated writes and restore tests.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
- Package or shard files appropriately and scale the supported metadata path while preserving data semantics.
- Set class-based retention and quotas while keeping legally required provenance and restorable checkpoints.
Explanation
Correct Answer
B:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Incorrect Answers
A:Use an approved durable checkpoint repository with atomic or validated writes and restore tests. That change would leave the reported problem unresolved: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
C:Package or shard files appropriately and scale the supported metadata path while preserving data semantics. That change would leave the reported problem unresolved: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
D:Set class-based retention and quotas while keeping legally required provenance and restorable checkpoints. That change would leave the reported problem unresolved: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Question 274
Two independent technical reviewers examine a AI Infrastructure Components and Architecture problem involving cooling and thermal headroom. Their shared incident log states: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. 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.
- Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
- Compare per-link utilization, incast windows, bisection demand, NIC speed and all-reduce job profiles during representative epochs.
- Reserve appropriate service classes and relieve burst contention after validating which path contributes to tail delay.
- Collect out-of-memory traces, per-rank allocator usage, peak activations, tensor sizes and chosen precision.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Explanation
Correct Answer
A:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
E:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
Incorrect Answers
B:Compare per-link utilization, incast windows, bisection demand, NIC speed and all-reduce job profiles during representative epochs. It would not provide the evidence needed for the paired diagnosis and follow-up: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
C:Reserve appropriate service classes and relieve burst contention after validating which path contributes to tail delay. It is not the scoped response supported by the accompanying evidence: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
D:Collect out-of-memory traces, per-rank allocator usage, peak activations, tensor sizes and chosen precision. It would not provide the evidence needed for the paired diagnosis and follow-up: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
Question 275
The acceptance board must decide whether an AI-infrastructure response to cooling and thermal headroom is fit for controlled operation. The fictional operating facts are: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. 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.
- GPU selection must consider model parameters, activation memory, optimizer states, precision and parallelism rather than nominal FLOPS.
- Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
- Revise supported precision, checkpointing or model sharding and select GPU memory capacity based on peak use.
- AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
- Multiple tenants operate within their allocations while a full-device benchmark remains unaffected.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Explanation
Correct Answer
B:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
D:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
F:A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Incorrect Answers
A:GPU selection must consider model parameters, activation memory, optimizer states, precision and parallelism rather than nominal FLOPS. It is not the rule that supports the proposed change and acceptance test: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
C:Revise supported precision, checkpointing or model sharding and select GPU memory capacity based on peak use. It does not implement the governing rule for this approval: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
E:Multiple tenants operate within their allocations while a full-device benchmark remains unaffected. It would not independently demonstrate that the control works: A GPU rack passes light tests but throttles during full-cluster training when hot-aisle recirculation raises inlet temperatures. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Question 276
A data-center AI architecture committee considers cooling and thermal headroom in its AI Infrastructure Components and Architecture review. At the affected site: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. 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.
- AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
- A low-average-latency network can still fail serving SLOs when transient queuing creates high p99 latency.
- Storage capacity plans require dataset lineage, scratch amplification, checkpoint retention and growth forecasts, not initial dataset size only.
- Distributed training requires adequate effective east-west throughput between GPU workers, with oversubscription based on real collective traffic.
Explanation
Correct Answer
A:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Incorrect Answers
B:A low-average-latency network can still fail serving SLOs when transient queuing creates high p99 latency. It does not govern the condition described: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
C:Storage capacity plans require dataset lineage, scratch amplification, checkpoint retention and growth forecasts, not initial dataset size only. It does not govern the condition described: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
D:Distributed training requires adequate effective east-west throughput between GPU workers, with oversubscription based on real collective traffic. It does not govern the condition described: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Question 277
An infrastructure diagnostician is investigating cooling and thermal headroom for the AI Infrastructure Components and Architecture team. The reported incident reads: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. 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.
- Map node-to-rack, PDU, network-fabric and storage failure domains alongside worker-placement requirements.
- Collect out-of-memory traces, per-rank allocator usage, peak activations, tensor sizes and chosen precision.
- Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
- Measure per-worker read concurrency, storage-node load, cache hit rates, queue depth and samples-per-second.
Explanation
Correct Answer
C:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Incorrect Answers
A:Map node-to-rack, PDU, network-fabric and storage failure domains alongside worker-placement requirements. That check would not establish the decisive facts for this condition: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
B:Collect out-of-memory traces, per-rank allocator usage, peak activations, tensor sizes and chosen precision. That check would not establish the decisive facts for this condition: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
D:Measure per-worker read concurrency, storage-node load, cache hit rates, queue depth and samples-per-second. That check would not establish the decisive facts for this condition: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Question 278
During a bounded AI-platform change-control review concerning cooling and thermal headroom, the site owner reports: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. 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.
- Size and balance the leaf-spine fabric for the observed communication pattern before adding accelerator nodes.
- Validate and stage a supported end-to-end software bill of materials with rollback images.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
- Use workload-aware anti-affinity and recovery checkpoints across independent domains where supported.
Explanation
Correct Answer
C:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
Incorrect Answers
A:Size and balance the leaf-spine fabric for the observed communication pattern before adding accelerator nodes. That change would leave the reported problem unresolved: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
B:Validate and stage a supported end-to-end software bill of materials with rollback images. That change would leave the reported problem unresolved: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
D:Use workload-aware anti-affinity and recovery checkpoints across independent domains where supported. That change would leave the reported problem unresolved: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Question 279
Two independent technical reviewers examine a AI Infrastructure Components and Architecture problem involving cooling and thermal headroom. Their shared incident log states: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. 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.
- Revise supported precision, checkpointing or model sharding and select GPU memory capacity based on peak use.
- Observe GPU idle gaps, CPU saturation, page faults, dataloader queue depth, NUMA affinity and PCIe locality.
- Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events.
- Review rack nameplate ratings, measured peak draw, dual-feed loading, breaker headroom and PSU redundancy.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Explanation
Correct Answer
C:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
E:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
Incorrect Answers
A:Revise supported precision, checkpointing or model sharding and select GPU memory capacity based on peak use. It is not the scoped response supported by the accompanying evidence: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
B:Observe GPU idle gaps, CPU saturation, page faults, dataloader queue depth, NUMA affinity and PCIe locality. It would not provide the evidence needed for the paired diagnosis and follow-up: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
D:Review rack nameplate ratings, measured peak draw, dual-feed loading, breaker headroom and PSU redundancy. It would not provide the evidence needed for the paired diagnosis and follow-up: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Compare inlet temperatures, coolant flow where relevant, facility heat rejection, exhaust recirculation and throttling events. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
Question 280
The acceptance board must decide whether an AI-infrastructure response to cooling and thermal headroom is fit for controlled operation. The fictional operating facts are: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. 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.
- AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
- Use workload-aware anti-affinity and recovery checkpoints across independent domains where supported.
- Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
- Serving capacity needs headroom for node failure, peak concurrency and warm model loading time.
- Training or serving completes a representative workload with safe high-water memory headroom.
- Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
Explanation
Correct Answer
A:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
C:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
F:An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads.
Incorrect Answers
B:Use workload-aware anti-affinity and recovery checkpoints across independent domains where supported. It does not implement the governing rule for this approval: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms.
D:Serving capacity needs headroom for node failure, peak concurrency and warm model loading time. It is not the rule that supports the proposed change and acceptance test: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. AI thermal designs must handle rack heat flux and liquid or air-cooling capacity under worst-case workloads. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.
E:Training or serving completes a representative workload with safe high-water memory headroom. It would not independently demonstrate that the control works: An edge AI lab triples accelerator density without verifying its building loop's heat-rejection capacity. Stress tests stay below equipment thermal limits with no recurring throttle or environmental alarms. Coordinate facility-validated cooling distribution and thermal controls before increasing sustained GPU power.