Answer-first summary for fast verification
Answer: Migrate the training jobs to Google Cloud TPUs to leverage their optimized performance for TensorFlow models.
Given the scenario, Cloud TPUs are the most effective solution. They are specifically designed to accelerate TensorFlow models, offering superior performance for large-scale matrix computations and big datasets. This choice aligns with the need for compliance, as Google Cloud ensures adherence to financial regulations, and provides scalability and cost-effectiveness by allowing dynamic resource allocation. While Nvidia GPUs are a viable alternative, they do not offer the same level of optimization for TensorFlow as TPUs. CPUs, whether Intel or AMD, are not suitable for such demanding tasks due to their lack of specialized hardware for deep learning workloads. For more details, refer to Google Cloud's TPU documentation.
Author: LeetQuiz Editorial Team
Ultimate access to all questions.
You are leading a project for a major banking institution that involves developing a deep neural network model using TensorFlow. The model's architecture is highly complex, requiring processing of massive datasets with intricate matrix computations. Currently, the training jobs are being executed on a cluster of virtual machines, but the duration of these jobs has extended to several weeks, threatening the project's deadline. The banking institution has emphasized the importance of compliance with financial regulations, cost-effectiveness, and the ability to scale resources dynamically based on the workload. Considering these constraints, what is the most effective solution to accelerate the training jobs while adhering to the project's requirements? (Choose one correct option)
A
Utilize Nvidia GPUs for their parallel processing capabilities.
B
Migrate the training jobs to Google Cloud TPUs to leverage their optimized performance for TensorFlow models.
C
Upgrade the virtual machines to use the latest Intel CPUs for improved computational efficiency.
D
Switch to AMD CPUs for their cost-effectiveness and energy efficiency.
No comments yet.