NCP-AII Question 87
Select 4You are tasked with validating the hardware installation of an NVIDIA DGX system in a data center. Which of the following steps are necessary to confirm that the hardware has been installed correctly and is ready for use?
- A
Verify that all GPUs are detected and functioning using the
nvidia-smicommand. - B
Ensure that the system's power cables are connected to a redundant power source.
- C
Run the
lscpucommand to confirm that the CPU has been overclocked for maximum performance. - D
Check that the system's cooling system is operational and maintaining optimal temperature levels.
- E
Validate the network connectivity by pinging the system's assigned IP address.
Show answer and explanation
Correct answers: A, B, D, E
Explanation
Validating hardware installation involves ensuring that all components (such as GPUs) are operational, power and cooling systems are functioning correctly, and the system is accessible on the network. These checks confirm that the system is ready for AI workloads. Overclocking the CPU, however, is not a standard validation step and can potentially harm system stability.
- A. Correct.
The
nvidia-smicommand is a key tool for verifying that the GPUs are properly installed and operational. It provides information on GPU status, utilization, and configurations. - B. Correct.
Ensuring the power cables are connected to a redundant power source is critical for maintaining system reliability and avoiding downtime in case of power failure.
- C. Incorrect.
Overclocking the CPU is not recommended for production systems like the NVIDIA DGX and is not part of standard hardware validation procedures.
- D. Correct.
A functional cooling system is essential for maintaining the thermal stability of the hardware, especially for high-performance computing systems like the DGX.
- E. Correct.
Validating network connectivity ensures that the system can communicate with other infrastructure components, which is crucial for AI workloads and distributed training.