Question 900
A project contributor checks the outcome after addressing this problem: A high-throughput AI data-preparation job receives a low-bandwidth storage path that stalls GPUs. Which result most directly confirms that the intended compute and AI workflow is working?
Explanation
Correct Answer
GPU idle time falls and data delivery meets the measured job throughput target. This outcome validates the targeted correction: Select and size a supported HPE storage arrangement using realistic read/write profiles. The original fault mechanism was: The data pipeline I/O rate cannot sustain the accelerator's consumption pattern.
Incorrect Answers
The data pipeline I/O rate cannot sustain the accelerator's consumption pattern. The proposed check does not demonstrate that this failure was corrected; it is intended for a condition in which the chosen placement offers no validated alternate capacity for service recovery. GPU idle time falls and data delivery meets the measured job throughput target.
Configuration deployment lacks staged validation and compatibility checks. The proposed result could verify work on that issue, but it would not establish success here. The data pipeline I/O rate cannot sustain the accelerator's consumption pattern. GPU idle time falls and data delivery meets the measured job throughput target.
The data pipeline I/O rate cannot sustain the accelerator's consumption pattern. The selected observation would assess a different failure: Shared datastore contention rather than CPU demand is limiting guest performance. Evidence for the current fix is: GPU idle time falls and data delivery meets the measured job throughput target.