Describe Compute Architectures Used with AI Workloads (CPU, GPU, TPU, FPGA) CERTIFIED-AI-EXPERT Practice Questions (Page 1)

Questions 1–5

1. 1expert · hard

   A cloud provider is designing a new instance type for AI inference. They want to support a wide variety of customer models, including both small recurrent neural networks and large transformer models. The instance must be cost-effective and allow customers to update their models frequently without changing hardware. Which architecture is most suitable for this flexible, multi-tenant inference service?

   • AAn FPGA
   • BA fixed ASIC
   • CA GPU
   • DA TPU
   Reveal Solution Discussion Select an answer first
2. 2expert · hard

   A company is designing a new AI accelerator for a specific deep learning model that will be deployed in millions of devices. The model is fixed and will not change. The company has a large budget for development and wants to minimize the per-device cost and power consumption. Which architecture is most appropriate?

   • AAn FPGA
   • BA custom ASIC
   • CA TPU
   • DA GPU
   Reveal Solution Discussion Select an answer first
4. 3application · medium

   A company runs a small AI service that performs occasional inference on a lightweight model. The workload is not performance-critical, and the team wants to minimize infrastructure cost. They already have underutilized servers with high-core-count CPUs. What is the most cost-effective approach?

   • APurchase a dedicated GPU server.
   • BDeploy on an FPGA.
   • CPurchase a TPU.
   • DDeploy the model on the existing CPU servers.
   Reveal Solution Discussion Select an answer first
5. 4foundation · easy

   Which of the following is a common application of GPUs in AI?

   • ARunning the control plane of a Kubernetes cluster
   • BStoring and retrieving large volumes of unstructured data
   • CTraining deep neural networks on large datasets
   • DCompiling source code into machine code
   Reveal Solution Discussion Select an answer first
6. 5foundation · easy

   For which type of AI task is an FPGA particularly well-suited?

   • ATraining large transformer models from scratch
   • BRunning a web server for an AI application
   • CLow-latency inference with a custom, optimized hardware pipeline
   • DGeneral-purpose data preprocessing on large datasets
   Reveal Solution Discussion Select an answer first