Develop Azure compute solutions →hardMultiple ChoiceObjective-mapped

AZ-204 Develop Azure compute solutions Practice Question

You are deploying a Docker container to Azure Container Instances (ACI). The container must use GPU resources for machine learning inference. You need to select the appropriate option to provision GPU-enabled containers. What should you do?

⚠ Common exam trap

Watch out — candidates often confuse local Docker GPU configuration (e.g., `--gpus all` in Dockerfile or docker run) with ACI's infrastructure-level GPU provisioning, assuming that a Dockerfile directive alone will enable GPU access in ACI, when in fact the SKU selection is mandatory and overrides any local settings.

Answer choices

Why each option matters

Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.

Correct answer & explanation

✓

Deploy the container to a container group with a GPU-enabled SKU (e.g., NV series).

Azure Container Instances supports GPU resources only when you deploy a container group using a GPU-optimized SKU, such as the NV-series (e.g., Standard_NC6s_v3). These SKUs provide NVIDIA Tesla GPUs (e.g., K80, P100, V100) that are directly exposed to the container, enabling hardware-accelerated machine learning inference. You must specify the GPU SKU in the container group's resource requests during deployment, and the container image must include the appropriate NVIDIA CUDA drivers or runtime.

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

✓
Deploy the container to a container group with a GPU-enabled SKU (e.g., NV series).
Why this is correct
To deploy a Docker container with GPU capabilities in Azure Container Instances, the correct approach is to provision a container group with a GPU-enabled SKU. This explicitly allocates a physical GPU from Azure's infrastructure, such as an NV-series VM, directly to your container instance. ACI then manages the necessary drivers and resources, making the GPU available to your containerized application for accelerated workloads like machine learning inference or training.
✗
Mount a GPU volume from the host.
Why it's wrong here
Azure Container Instances operates as a fully managed service, abstracting away the underlying host infrastructure. Consequently, ACI does not support directly mounting GPU devices or volumes from the host system, unlike a self-managed Docker environment. GPU resources are instead provisioned and exposed through the container group's selected SKU, not by direct device mapping or host volume mounts.
✗
Use Azure Batch with GPU-enabled pools.
Why it's wrong here
Azure Batch is a distinct Azure service designed for large-scale parallel and high-performance computing workloads, often utilizing GPU-enabled virtual machines within its compute pools. While capable of running GPU-accelerated tasks, it is not Azure Container Instances. The question specifically targets ACI, making Azure Batch an incorrect solution for deploying a container *to ACI*.
✗
Enable container GPU support in the Dockerfile.
Why it's wrong here
A Dockerfile defines the build instructions and environment for a container image, including installing GPU-aware libraries or CUDA toolkits. However, a Dockerfile cannot provision or enable the underlying physical GPU hardware itself. The actual GPU resource allocation is an infrastructure concern, which in ACI is handled by selecting a specific GPU-enabled container group SKU during deployment, not by image definition.

Go deeper

AZ-204 Develop Azure compute solutions Practice Question

Why each option matters

Option-by-option breakdown

Related to this question