NVIDIA AI Infrastructure - NCP-AII Exam Practice Test

A. Run DCGM diagnostics alongside burn-in to monitor GPU health metrics

B. Reduce blocksize to 500MB to lower memory pressure

C. Switch from BERT to GPT models for simpler computations

D. Enable HPL_USE_NVSHMEM for alternative memory sharing

A. Perform a cold reset, stop all GPU activity, update and reboot the BMC, update motherboard and tray components, and verify completion.

B. Stop all GPU activity, update and reboot the BMC, update motherboard and tray components, perform a cold reset, and verify completion.

C. Update the BMC and skip the GPU tray and motherboard tray updates if the system appears healthy.

D. Update the GPU tray first, then the motherboard tray, and reboot the BMC after all updates are complete.

A. Frequent peer-to-peer communication and collective operations among GPUs

B. Local NVMe read performance for training datasets

C. Kubernetes control plane scheduling latency

D. Sequential CPU-intensive preprocessing before model execution

A. ethtool

B. iblinkinfo

C. mlxconfig

D. flint

A. Validate that the power distribution system supports N+1 redundancy with sufficient capacity for peak loads

B. Confirm that the power supply units (PSUs) in each DGX H100 system are rated for 80 PLUS Gold efficiency or higher.

C. Ensure that each rack has at least one power circuit capable of delivering 10.2 kW of power.

D. Verify the input voltage range (200-240V AC) for compatibility with the data center's power infrastructure.

A. Reinstall drivers using apt-get install nvidia-driver-550 without rebooting.

B. Enroll the Machine Owner Key (MOK) during system reboot and enter the recorded password.

C. Delete /etc/X11/xorg.conf to force driver reconfiguration.

D. Disable Secure Boot permanently in BIOS/UEFI settings.

A. broadcast_perf -b 8 -e 16G -f2 -g 8 without split configuration

B. all_reduce_perf -b 8 -e 16G -f2 -g 8 with NCCL_TESTS_SPLIT= "OR 0x7"

C. all_reduce_perf -b 8 -e 16G -f2 -g 4 with NCCL_TESTS_SPLIT= "MOD 2"

D. all_reduce_perf -b 8 -e 16G -f2 -g 1 repeated 8 times

A. Hardware acceleration for tensor operations during model training

B. Automatic GPU virtualization through CUDA

C. Infrastructure offload with isolated security and networking services on the DPU

D. Dynamic expansion of GPU framebuffer memory

A. Perform a cold reset, stop all GPU activity, update and reboot the BMC, update motherboard and tray components, and verify completion.

B. Stop all GPU activity, update and reboot the BMC, update motherboard and tray components, perform a cold reset, and verify completion.

C. Update the BMC and skip the GPU tray and motherboard tray updates if the system appears healthy.

D. Update the GPU tray first, then the motherboard tray, and reboot the BMC after all updates are complete.

A. Begin validation by pinging each node's management interface and confirming the network cabling is correct when hosts respond to management pings.

B. Run the cable validation process incrementally during deployment (section by section), to catch and resolve errors as early as possible.

C. Focus on validating the highest bandwidth links in the design first, deferring non-critical cable mislabelings to be fixed after initial workloads are deployed and tested.

D. Compare every cable's physical connection to the planned topology diagram and validate correct ports and link paths.

A. NVSwitch enables high-bandwidth, low-latency communication among all GPUs within the server.

B. PCIe automatically aggregates bandwidth across every GPU.

C. NVSwitch accelerates Ethernet routing between servers.

D. NVLink replaces system memory with GPU memory.

A. Start by issuing nvidia-smi -L to list GPUs, followed by nvsm --refresh to clear all alerts, and nvidia-smi -q to get detailed GPU information.

B. Use nvsm show health for a system health summary, nvsm show storage for storage issues, and nvidia-smi -q to get detailed GPU information.

C. Run nvsm collect-stats to gather logs, use lsblk to understand if there are storage problems, and nvidia-smi -q to get detailed GPU information.

D. Run nvsm reset to restore system health, then use nvidia-smi --fix for automatic GPU repairs and status recovery.

A. Export all switch logs and grep for "FW Version"

B. Run mlxlink -d lid-<LID> -m on each port manually

C. Use "Topology" view to visually inspect cable icons.

D. Navigate to "Devices" > select a switch > "Cables" tab to see ASIC firmware and transceiver versions.