You are a data scientist analyzing a social media network with NVIDIA cuGraph to identify the most influential users using the PageRank algorithm.
Which option best describes how cuGraph PageRank operates on a directed graph?

• A. PageRank assigns equal importance to all nodes in the graph initially and updates values only based on outgoing edges, ignoring incoming edges.
• B. PageRank in cuGraph operates only on undirected graphs and cannot be applied to networks where edges have a direction.
• C. PageRank in cuGraph uses an iterative power method to update node importance values based on incoming edges, incorporating a damping factor to handle random jumps.
• D. PageRank in cuGraph is a label propagation algorithm that clusters nodes into communities rather than ranking their importance.

You need to train a deep learning model using PyTorch on a dataset too large for a single GPU. You decide to use Dask with NVIDIA GPUs for multi-GPU scaling.
Which approach is the most effective for distributing the workload?

• A. Use Dask.delayed to wrap PyTorch training functions and schedule them across multiple GPUs
• B. Use Dask Bag to shard the dataset and train separate PyTorch models on each shard
• C. Use Dask-CUDA workers with PyTorch's DistributedDataParallel (DDP) for training across multiple GPUs
• D. Use Dask's built-in deep learning API to automatically distribute PyTorch models across GPUs

You are training a deep learning model on a large dataset and are deciding whether to use a single GPU or multiple GPUs.
Which of the following are true considerations when comparing single-GPU and multi-GPU training setups? (Select two)

• A. Multi-GPU setups perform better only when the batch size is reduced.
• B. Single-GPU training is limited by the VRAM (video memory) on the GPU, so larger models or datasets may require multi-GPU setups.
• C. Single-GPU training is generally more cost-effective and should be preferred unless scaling is absolutely necessary.
• D. Multi-GPU training can significantly reduce training time when the dataset is large and the model is computationally intensive.
• E. Multi-GPU training requires modifications to the model architecture to make it compatible with parallel processing.

Which of the following statements best describes the role of GPUs in accelerating data science workloads?

• A. GPUs are only effective for acceleration when used in conjunction with Tensor Processing Units (TPUs), as they cannot train deep learning models independently.
• B. GPUs are designed primarily for rendering graphics and have limited utility in machine learning and deep learning applications.
• C. GPUs are optimized for sequential data processing tasks, making them more efficient than CPUs for database operations.
• D. GPUs use thousands of smaller cores that can execute many parallel computations simultaneously, making them ideal for large-scale matrix operations.