Medium1 markMultiple Choice

GCP ACE · Question 21 · Domain 3.1: Deploying and implementing Compute Engine resources

You are deploying a machine learning model on Compute Engine that requires GPU acceleration. You have created a VM and attached an NVIDIA Tesla T4 GPU.

Which TWO additional steps MUST you perform before your application can utilize the GPU? (Select TWO)

Answer options:

A.

Install the NVIDIA GPU drivers on the guest OS.

B.

Enable the Cloud TPU API in the project.

C.

Install the CUDA toolkit/libraries on the guest OS.

D.

Configure a local SSD to act as VRAM for the GPU.

E.

Assign the 'GPU Admin' IAM role to the Compute Engine service account.

How to approach this question

Identify the software requirements for utilizing NVIDIA GPUs inside a VM.

Full Answer

When you attach a GPU to a Compute Engine instance, it is presented as hardware to the guest OS. To use it, you must install the NVIDIA device drivers. For machine learning workloads, you also need to install the CUDA toolkit/libraries, which provide the programming interface for the GPU.

Common mistakes

Assuming GCP automatically installs drivers on all OS images (only specific Deep Learning VM images have them pre-installed).

Practice the full GCP Associate Cloud Engineer Practice Exam 4

50 questions · hints · full answers · grading

More questions from this exam