Integrate Compute for GPU Nodes into an Existing System Architecture POWEREDGE-XE-OPERATE Practice Questions (Page 3)

Questions 11–15

1. 11expert · hard

   A team has installed a Dell PowerEdge XE9680 with NVIDIA H100 GPUs into an existing RHEL 9 cluster. After installing the NVIDIA driver, `nvidia-smi` shows the GPUs, but the system log reports 'NVML_SUCCESS' but the GPUs are not being used by the container runtime. The administrator has confirmed that the driver is loaded and the GPUs are visible. What is the most likely cause?

   • AThe GPU node's firmware is outdated
   • BThe GPU node's PCIe link is running at a reduced speed
   • CThe container runtime (e.g., Docker or Podman) is not configured with the NVIDIA Container Toolkit
   • DThe GPU node's power supply is insufficient
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   Which maintenance task is recommended to prevent GPU nodes from failing due to dust accumulation?

   • ADisabling the GPU's power management features
   • BUpdating the GPU firmware frequently
   • CIncreasing the GPU's memory allocation
   • DRegularly cleaning the cooling fans and air filters
   Reveal Solution Discussion Select an answer first
4. 13foundation · easy

   Which type of workload is best suited for GPU nodes rather than traditional CPU-only compute resources?

   • ALow-latency database transactions
   • BSimple sequential tasks like basic file serving
   • CHighly parallel tasks such as machine learning training or scientific simulations
   • DNetwork routing and firewall filtering
   Reveal Solution Discussion Select an answer first
5. 14foundation · easy

   When planning for scalability of GPU nodes, which factor is most important to consider to accommodate future expansion?

   • AThe number of USB ports on each node
   • BThe availability of additional power and cooling capacity in the data center
   • CThe color of the server racks
   • DThe type of keyboard used for administration
   Reveal Solution Discussion Select an answer first
6. 15foundation · easy

   What is the primary purpose of implementing a management solution for GPU nodes in an existing system architecture?

   • ATo automatically upgrade the GPU hardware
   • BTo increase the GPU nodes' processing speed
   • CTo replace the need for physical security measures
   • DTo monitor and control the GPU nodes' health and performance
   Reveal Solution Discussion Select an answer first