Hard1 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 and need to attach NVIDIA GPUs to the instance to accelerate training.

Which TWO requirements must be met to successfully use GPUs on Compute Engine? (Select TWO)

Answer options:

A.

The instance must not be a shared-core machine type (e.g., e2-micro).

B.

You must install the appropriate NVIDIA drivers and CUDA libraries on the guest OS.

C.

You must use a Windows Server operating system.

D.

The instance must be placed in a region that supports Cloud TPU.

E.

You must enable the 'GPU API' in the API library.

How to approach this question

Recall the hardware and software prerequisites for attaching GPUs to GCE instances.

Full Answer

The instance must not be a shared-core machine type (e.g., e2-micro)., You must install the appropriate NVIDIA drivers and CUDA libraries on the guest OS.
To use GPUs on Compute Engine, you must attach them to a supported machine type (shared-core machines like e2-micro are not supported). Furthermore, simply attaching the virtual hardware is not enough; you must install the NVIDIA device drivers and CUDA libraries on the guest operating system so the software can interface with the GPU.

Common mistakes

Forgetting that drivers must be installed manually (or via a startup script/custom image) after the VM is provisioned.

Practice the full GCP Associate Cloud Engineer Practice Exam 7

50 questions · hints · full answers · grading

More questions from this exam