NCP-AIO Workload Management Practice Question
A research team is submitting many short-lived experiment jobs to an NVIDIA-accelerated Kubernetes cluster. The operations team wants to reduce GPU idle time and improve overall utilization without modifying the training code. Which TWO approaches should the operations team implement? (Choose two.)
⚠ Common exam trap
The trap here is assuming that dedicating a GPU per short job or requesting more GPUs improves utilization, when in fact both reduce concurrency and leave GPUs idle.
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
✓
Use a gang-scheduling or batch scheduler with backfill so queued jobs start as soon as GPUs free up
Short-lived experiment jobs create idle GPU gaps. Time-slicing allows multiple such pods to share a device, and a batch scheduler with backfill keeps the queue moving so GPUs are assigned as soon as capacity frees. Both measures raise utilization without touching training code, which is exactly what the operations team needs.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Pin each experiment job to a dedicated physical GPU using node affinity
Why it's wrong here
Dedicating a physical GPU to each short-lived job prevents sharing and leaves the GPU idle whenever a job finishes or waits on data loading. This approach lowers utilization rather than improving it, and it does not address the intermittent idle periods that occur between short experiment runs.
- ✓
Use a gang-scheduling or batch scheduler with backfill so queued jobs start as soon as GPUs free up
Why this is correct
Batch schedulers with backfill can start smaller jobs ahead of larger queued jobs when resources are available, keeping GPUs busy between experiment runs. This reduces idle gaps and improves throughput for many short-lived jobs, complementing GPU sharing without requiring modifications to the training scripts.
- ✗
Disable the NVIDIA device plugin and mount GPUs directly via hostPath
Why it's wrong here
Bypassing the device plugin removes Kubernetes' ability to track and schedule GPU resources, leading to conflicts and unpredictable placement. It does not improve utilization; instead it undermines scheduling accuracy and can cause multiple pods to unknowingly contend for the same device, which is worse than the current situation.
- ✗
Increase the pod's nvidia.com/gpu request to reserve more GPU memory
Why it's wrong here
Requesting more GPUs reserves additional devices and reduces the number of jobs that can run concurrently. For short experiments that do not need multiple GPUs, this wastes capacity and increases idle time, moving utilization in the wrong direction and contradicting the stated objective.
- ✓
Enable GPU sharing with time-slicing so multiple experiment pods can occupy the same GPU
Why this is correct
Time-slicing lets the device plugin advertise multiple virtual GPU replicas per physical device, so several short experiment pods can run concurrently on one GPU. This reduces idle time between jobs and raises utilization without any change to the training code, directly addressing the goal of the scenario.
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.