NCP-AII Question 68
Select 3You are tasked with installing multiple NVIDIA GPUs into a new AI server. Which of the following steps should you take to ensure the GPUs are installed properly and function effectively?
- A
Ensure the server motherboard has sufficient PCIe slots and compatible power connectors for the GPUs.
- B
Install the GPUs without checking the server's cooling system, as modern GPUs have built-in cooling mechanisms.
- C
Verify that the power supply unit (PSU) can handle the combined power requirements of all GPUs and other server components.
- D
Update the server BIOS/UEFI firmware to the latest version to ensure compatibility with the installed GPUs.
- E
Skip grounding yourself before handling the GPUs, as modern hardware is resistant to electrostatic discharge (ESD).
Show answer and explanation
Correct answers: A, C, D
Explanation
When installing physical GPUs, it is important to ensure compatibility with the server's motherboard and power supply, verify sufficient cooling capabilities, and update firmware to avoid hardware or software compatibility issues. Skipping grounding procedures or neglecting cooling measures can result in hardware damage or suboptimal performance.
- A. Correct.
Correct: Ensuring the motherboard has sufficient PCIe slots and adequate power connectors is a critical first step to ensure the GPUs can be physically installed and powered.
- B. Incorrect.
Incorrect: Installing GPUs without verifying the cooling system is a mistake, as GPUs generate significant heat during operation, and inadequate cooling can lead to performance throttling or hardware damage.
- C. Correct.
Correct: Verifying the PSU's capacity to handle the GPUs' power requirements is essential to avoid power shortages or instability during operation.
- D. Correct.
Correct: Updating the BIOS/UEFI firmware is necessary to ensure the system can recognize and properly interface with the installed GPUs.
- E. Incorrect.
Incorrect: Skipping grounding before handling GPUs can result in electrostatic discharge (ESD) damage, which can compromise or permanently damage sensitive hardware components.