NCP-AII Question 63
Select 3You are tasked with installing a GPU-based server for an AI workload. After physically installing the GPUs and booting the system, which of the following steps should you perform to ensure the GPUs are properly configured and usable for workloads?
- A
Install the appropriate NVIDIA GPU drivers for the installed GPUs.
- B
Verify the GPUs are detected using a tool such as
nvidia-smi. - C
Configure the server’s BIOS to enable GPU passthrough for virtualized environments.
- D
Ensure the CUDA Toolkit is correctly installed if CUDA-based applications are required.
- E
Update the server's kernel to remove any version compatibility issues with the GPUs.
Show answer and explanation
Correct answers: A, B, D
Explanation
Installing GPU-based servers involves several critical steps after the physical installation, including installing the appropriate drivers, verifying GPU detection, and setting up the CUDA Toolkit if required. These steps ensure that the GPUs are correctly configured and ready for AI workloads. However, not all steps, such as enabling GPU passthrough or updating the kernel, are universally necessary and depend on the specific use case or environment.
- A. Correct.
Correct. Installing the appropriate NVIDIA GPU drivers is essential for the server to recognize and leverage the GPUs.
- B. Correct.
Correct. Verifying GPU detection using tools like
nvidia-smiensures that the system recognizes the installed GPUs and they are functioning as expected. - C. Incorrect.
Incorrect. GPU passthrough is only required if you are planning to use the GPUs in a virtualized environment, which is not a necessary step for all installations.
- D. Correct.
Correct. The CUDA Toolkit is crucial if you are running CUDA-based applications, as it provides the necessary libraries and tools.
- E. Incorrect.
Incorrect. While keeping the kernel updated is good practice, it is not specifically required in this scenario unless there is a known incompatibility issue.