Practice Free NCA-AIIO Exam Online Questions
What is an advantage of InfiniBand over Ethernet?
- A . InfiniBand always provides higher bandwidth than Ethernet.
- B . InfiniBand supports RDMA while Ethernet does not.
- C . InfiniBand offers lower latency than Ethernet.
C
Explanation:
InfiniBand’s advantage over Ethernet lies in its lower latency, achieved through a streamlined protocol and hardware offloads, delivering microsecond-scale communication critical for AI clusters. While InfiniBand often offers high bandwidth, Ethernet can match or exceed it (e.g., 400 GbE), and Ethernet supports RDMA via RoCE, making latency the standout differentiator. (Reference: NVIDIA Networking Documentation, Section on InfiniBand vs. Ethernet)
What is the maximum number of MIG instances that an H100 GPU provides?
- A . 7
- B . 8
- C . 4
A
Explanation:
The NVIDIA H100 GPU supports up to 7 Multi-Instance GPU (MIG) partitions, allowing it to be divided into seven isolated instances for multi-tenant or mixed workloads. This capability leverages the H100’s architecture to maximize resource flexibility and efficiency, with 7 being the documented maximum.
(Reference: NVIDIA H100 GPU Documentation, MIG Section)
A data science team compares two regression models for predicting housing prices. Model X has an R-squared value of 0.85, while Model Y has an R-squared value of 0.78. However, Model Y has a lower Mean Absolute Error (MAE) than Model X.
Based on these statistical performance metrics, which model should be chosen for deployment, and why?
- A . Model X should be chosen because it is likely to perform better on unseen data.
- B . Model X should be chosen because a higher R-squared value indicates it explains more variance in the data.
- C . Model Y should be chosen because a lower MAE indicates it has better prediction accuracy.
- D . Model X should be chosen because R-squared is a more comprehensive metric than MAE.
A data scientist is working on a research project that involves identifying factors affecting product sales using a large dataset. They aim to uncover hidden relationships and trends by leveraging NVIDIA GPUs for data processing.
Which two methods should they employ to ensure the results are both accurate and meaningful? (Select two)
- A . Apply correlation matrix analysis to identify relationships between variables.
- B . Use a simple random sampling technique to analyze only a subset of the data.
- C . Focus on univariate analysis to reduce complexity.
- D . Conduct time series analysis on the sales data to identify trends over time.
- E . Ignore multicollinearity between variables to simplify the model.
Your team is tasked with accelerating a large-scale deep learning training job that involves processing a vast amount of data with complex matrix operations. The current setup uses high-performance CPUs, but the training time is still significant.
Which architectural feature of GPUs makes them more suitable
than CPUs for this task?
- A . Low power consumption
- B . Large cache memory
- C . Massive parallelism with thousands of cores
- D . High core clock speed
You have developed two different machine learning models to predict house prices based on various features like location, size, and number of bedrooms. Model A uses a linear regression approach, while Model B uses a random forest algorithm. You need to compare the performance of these models to determine which one is better for deployment.
Which two statistical performance metrics would be most appropriate to compare the accuracy and reliability of these models? (Select two)
- A . Mean Absolute Error (MAE)
- B . Cross-Entropy Loss
- C . F1 Score
- D . R-squared (Coefficient of Determination)
- E . Learning Rate
You are tasked with designing an AI system that needs to handle both training and inference workloads. The system must optimize performance while minimizing costs, given that training is done periodically, but inference is performed continuously in real-time.
What considerations should be made in designing the architecture to meet both training and inference requirements efficiently? (Select two)
- A . Using a mixed-precision approach for both training and inference.
- B . Ensuring redundancy and failover for inference systems.
- C . Deploying GPUs with more memory for inference than for training.
- D . Allocating more GPUs for inference compared to training.
- E . Prioritizing high-bandwidth interconnects between GPUs for training.
You are managing a multi-node AI cluster that performs real-time analytics on streaming data from financial markets. The data is processed in parallel across multiple GPUs within a high-performance computing (HPC) environment. Recently, the cluster has been experiencing delays in aggregating the final results, causing the analytics to fall behind market events.
Which action would MOST likely resolve the delay in aggregating results in this HPC environment?
- A . Implementing more aggressive data compression before transmission between nodes.
- B . Increasing the number of GPUs in each node to improve processing speed.
- C . Switching to a batch processing model for handling data streams.
- D . Optimizing the network fabric to reduce latency between the nodes.
You are comparing two regression models, Model X and Model Y, that predict stock prices. Model X has an R-squared (proportion of explained variance) of 0.75, while Model Y has an R-squared of 0.85.
Which model should you prefer based on the R-squared metric, and what does this metric indicate about the model’s performance?
- A . Model X is better because a lower R-squared indicates more flexibility.
- B . Neither model is better because R-squared is not a reliable metric.
- C . Model Y is better because it has a higher R-squared value, indicating it explains more variance in the data.
- D . Model X is better because it might generalize better despite a lower R-squared.
You are working with a team of data scientists on an AI project where multiple machine learning models are being trained to predict customer churn. The models are evaluated based on the Mean Squared Error (MSE) as the loss function. However, one model consistently shows a higher MSE despite having a more complex architecture compared to simpler models.
What is the most likely reason for the higher MSE in the more complex model?
- A . Low learning rate in model training
- B . Overfitting to the training data
- C . Incorrect calculation of the loss function
- D . Underfitting due to insufficient model complexity