NCP-AIO Troubleshooting and Optimization Practice Question
An administrator wants to ensure that a training process is limited to a single GPU on a multi-GPU node. Which environment variable should be set?
⚠ Common exam trap
Students often mistakenly select command-line flags or code-level device placement arguments instead of the standard operating system environment variable required to restrict GPU visibility globally.
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
✓
CUDA_VISIBLE_DEVICES=0
Controlling GPU visibility is a fundamental skill for resource management in multi-tenant environments. By using CUDA_VISIBLE_DEVICES, an administrator can restrict a process to a specific device, preventing multiple jobs from competing for the same GPU. This isolation is crucial for maintaining performance stability and ensuring that individual jobs receive consistent, predictable access to compute resources without interference from other concurrent tasks.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
NCCL_DEBUG=INFO
Why it's wrong here
NCCL_DEBUG is an environment variable used to increase the verbosity of the NCCL collective communication library for troubleshooting purposes. It does not influence the GPU visibility or the physical mapping of processes to specific GPU devices within a multi-GPU server environment.
- ✓
CUDA_VISIBLE_DEVICES=0
Why this is correct
CUDA_VISIBLE_DEVICES is the standard environment variable used to mask specific GPUs from a process. Setting it to a specific index restricts the application to use only that hardware device, which is the standard method for isolating jobs in a multi-GPU system.
- ✗
NVIDIA_DRIVER_CAPABILITIES=compute
Why it's wrong here
This variable is used when launching containers to specify the driver capabilities available to the container, such as graphics or compute. It does not control the selection or restriction of specific GPU devices, which is handled via the CUDA runtime interface and environment variables.
- ✗
OMP_NUM_THREADS=1
Why it's wrong here
OMP_NUM_THREADS controls the number of threads used by OpenMP for CPU-based parallel computation. It has no effect on the GPU hardware allocation or visibility. While it is important for performance, it cannot be used to restrict a process to a specific GPU device.
About these practice questions
One of 309 original NCP-AIO practice questions on Courseiva, each with a full explanation and wrong-answer analysis — not exam dumps or protected exam content. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official NVIDIA exam blueprint
This NCP-AIO practice question is part of Courseiva's free NVIDIA certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the NCP-AIO exam.