Free NVIDIA NCP-IB Exam Questions

Question 1

You are responsible for the deployment of a large-scale data center using NVIDIA's InfiniBand for AI workloads. Your goal is to ensure both the scalability and reliability of the fabric as the data center grows over time. The initial deployment involves 100 nodes, but you plan to expand to 500 nodes within the next year. It's critical that performance scales with the growth, minimizing bottlenecks, and ensuring reliability without network downtime. Which best practice should you implement to ensure scalability and reliability in a production InfiniBand environment as your data center grows?

A : Implement a spine-leaf topology using adaptive routing and a lossless network design.

B : Deploy a multi-tier fat-tree topology with oversubscription to optimize cost and limit hardware.

C : Use a single large switch for all nodes to reduce latency and simplify network management.

D : Enable Flow Control on all ports and restrict packet sizes to prevent congestion.

Answer: A

Question 2

You are configuring the HCAs in your InfiniBand environment to maximize performance for a high-throughput, low-latency application. Which tuning parameter should you adjust on the HCA to reduce latency without causing packet loss?

A : Increase the MTU (Maximum Transmission Unit) to its highest supported value.

B : Decrease the MTU (Maximum Transmission Unit) to minimize the time it takes for small packets to be transmitted.

C : Adjust the Receive Queue Depth (RQD) to increase the ability to handle incoming traffic bursts.

D : Tune the SL (Service Level) to prioritize the application traffic over other traffic.

Answer: C

Question 3

You are configuring an InfiniBand-based high-performance computing (HPC) cluster for a scientific simulation workload that requires low latency and high throughput communication between nodes. The application is designed to leverage RDMA (Remote Direct Memory Access) to bypass the CPU and improve data transfer efficiency. You notice that while the latency is low, the throughput is not meeting expected levels. Which of the following is the most effective method to optimize RDMA performance for this application?

A : Enable CPU coalescing to allow RDMA to use more CPU resources

B : Ensure proper alignment of memory buffers to cache lines.

C : Increase MTU size on InfiniBand network interfaces

D : Disable flow control mechanisms for RDMA transfers

Answer: B

Question 4

You are configuring an InfiniBand-based high-performance computing (HPC) cluster for a scientific simulation workload that requires low latency and high throughput communication between nodes. The application is designed to leverage RDMA (Remote Direct Memory Access) to bypass the CPU and improve data transfer efficiency. You notice that while the latency is low, the throughput is not meeting expected levels. Which of the following is the most effective method to optimize RDMA performance for this application?

A : Enable CPU coalescing to allow RDMA to use more CPU resources

B : Ensure proper alignment of memory buffers to cache lines.

C : Increase MTU size on InfiniBand network interfaces

D : Disable flow control mechanisms for RDMA transfers

Answer: B

Question 5

How does Partitioning in an InfiniBand network enhance security, and what is the correct method for configuring it to prevent unauthorized communication between devices?

A : Partitioning configures access control lists based on device MAC addresses, preventing unauthorized devices from joining the fabric.

B : Partitioning isolates traffic by assigning devices to specific Partition Keys (P_Keys), and only devices with matching P_Keys can communicate, providing strong segmentation.

C : Partitioning uses VLAN tags similar to Ethernet networks, ensuring that only devices within the same VLAN can communicate, thereby preventing unauthorized traffic.

D : Partitioning creates encrypted tunnels between InfiniBand devices, ensuring secure communication even if unauthorized devices attempt to intercept the traffic.

Answer: B