Question 1

After running a 24-hour stress test on a DGX node, the administrator should verify which two keymetrics to ensure system stability?

Options :

A : Average CPU usage >80% and Docker container uptime.

B : No thermal throttling events and consistent GPU utilization >95% throughout the test.

C : SSD write endurance and RAM capacity.

D : Total energy consumption and NVLink bandwidth.

Answer: B

Question 2

An engineer needs to verify the current firmware versions of all components (ATF, BSP, NIC, UEFI) ona BlueField-3 DPU's BMC. Which Redfish API command provides this information?

Options :

A :

B :

C :

D :

redfish/v1/UpdateService/FirmwareInventory

Answer: D

Question 3

During East-West fabric validation on a 64-GPU cluster, an engineer runs all_reduce_perf andobserves an algorithm bandwidth of 350 GB/s and bus bandwidth of 656 GB/s. What does thisindicate about the fabric performance?

Options :

A : Inconclusive; rerun with point-to-point tests.

B : Optimal performance; bus bandwidth near theoretical peak for NDR InfiniBand.

C : Critical failure; bus bandwidth exceeds hardware capabilities.

D : Suboptimal performance; algorithm bandwidth should match bus bandwidth.

Answer: B

Question 4

After NCCL burn-in reports "transport retry count exceeded," which corrective action addresses theunderlying fabric issue?

Options :

A : Switch from Ring to Tree algorithms via NCCL_ALGO=TREE

B : Reduce message size to decrease network utilization

C : Increase NCCL_IB_TIMEOUT to tolerate longer latencies

D : Inspect InfiniBand link quality metrics (BER, symbol errors) and replace faulty cables

Answer: D

Question 5

For an NVIDIA Enterprise AI Factory with 256 GPUs, which storage solution characteristic is mostcritical to validate during scaling tests?

Options :

A : Consistent per-node throughput >8 GiB/s.

B : Single-node write performance during idle clusters.

C : RAID rebuild times under disk failure.

D : Maximum 4K random read IOPS exceeding 1 million.

Answer: A