Important NVIDIA NCP-AII Exam Questions

Answer: C

Explanation:

NVIDIA Multi-Instance GPU (MIG) technology, introduced with the Ampere architecture (A100) and enhanced in Hopper (H100), allows a single physical GPU to be partitioned into multiple isolated instances. To manage these instances and their scheduling behavior (such as moving from a default time-sliced scheduler to a fixed-share or 'MIG' partitioned mode), the nvidia-smi utility is used. The command nvidia-smi -i 0 -mig 1 enables MIG mode on the first GPU (index 0). Once MIG is enabled, the administrator can create specific GPU instances with dedicated compute and memory resources. This 'Fixed Share' approach ensures that one tenant's workload does not impact the performance or latency of another, which is critical for deterministic AI inference and multi-tenant cloud environments. Option A and B refer to VMware ESXi specific commands which are not the primary method for raw hardware configuration in standard AI infrastructure, and Option D is for network adapter configuration.

Answer: B

Explanation:

HPL-AI (High-Performance Linpack for AI) differs from traditional HPL by utilizing lower-precision arithmetic (FP16/BF16/TF32) while maintaining FP64 accuracy through iterative refinement. The significant jump in performance seen in version 2.0 and above is largely attributed to advancements in the communication layer. In multi-node DGX clusters, the CPU often becomes a bottleneck when managing MPI (Message Passing Interface) ranks and coordinating data transfers between GPUs across the InfiniBand fabric. By implementing MPI-aware GPU communication (and leveraging technologies like SharpV3 or NCCL integration), the benchmark significantly reduces the 'GPU Idle Time' spent waiting for the CPU to orchestrate the next computation block. This optimization ensures that the GPUs remain at near 100% utilization during the intensive matrix multiplication phases. By offloading collective operations and using GPUDirect RDMA, the system bypasses the host memory and CPU, effectively doubling the effective throughput of the benchmark compared to older, CPU-heavy coordination methods.

Answer: A

Explanation:

When deploying NVIDIA vGPU on VMware ESXi, the NVIDIA driver provides several scheduling policies to determine how GPU physical resources are shared among multiple virtual machines. The default behavior is often the 'Best Effort' scheduler, but for environments requiring predictable performance across all users, the 'Equal Share' scheduler is preferred. This scheduler gives each vGPU an equal 'time slice' of the physical GPU's engines. The configuration is managed via module parameters passed to the nvidia kernel driver during host boot. The specific registry key for this behavior is RmPVMRL. Setting RmPVMRL=0x01 enables the Equal Share scheduler (Option A). Conversely, 0x00 would revert to the default time-sliced behavior. It is critical to use system module parameters set to ensure the setting persists across reboots and is applied globally to the NVIDIA driver stack. This ensures that no single 'noisy neighbor' VM can monopolize the GPU cycles, which is a common requirement in shared AI research labs or virtual desktop infrastructures where consistency is more important than raw peak throughput of a single task.

Answer: A

Explanation:

High-Performance Linpack (HPL) is a memory-intensive benchmark that allocates a large portion of available GPU memory to store the matrix $N$. While a server may have 2TB of physical system RAM, the 'not enough memory' error usually refers to the HBM (High Bandwidth Memory) on the GPUs themselves. In a DGX H100 system, each GPU has 80GB of HBM3. If the problem size ($N$) specified in the HPL.dat file is too large, the required memory for the matrix will exceed the aggregate capacity of the GPU memory. Reducing the problem size ($N$) while maintaining the optimal block size ($NB$) ensures that the problem fits within the GPU memory limits while still pushing the computational units to their peak performance. Increasing the block size (Option C) would actually increase the memory footprint of certain internal buffers, potentially worsening the issue. Reducing $N$ is the standard procedure to stabilize the run during the initial tuning phase of an AI cluster bring-up.

Answer: A

Explanation:

Interpreting NCCL (NVIDIA Collective Communications Library) results requires understanding the difference between theoretical link rate and effective bus bandwidth. A 400G (NDR) InfiniBand link has a theoretical unidirectional capacity of 50 GB/s. In an 8-GPU DGX system, where each GPU is typically mapped to its own 400G HCA, the aggregate theoretical bandwidth is 400 GB/s. However, NCCL performance benchmarks report Bus Bandwidth, which accounts for the mathematical overhead of collective operations like all_reduce. Achieving 350 GB/s (Option A) is widely considered optimal and representative of a healthy, well-configured fabric. This indicates that the system is effectively utilizing GPUDirect RDMA to bypass host memory and that the network switches are routing traffic without significant congestion or packet drops. Expecting >390 GB/s (Option C) is unrealistic once protocol headers and algorithmic overhead are subtracted from the raw bit rate. Rerunning with CPU stress (Option D) would not provide insights into the InfiniBand/GPU fabric health.