Characterization of high-resolution AI data center training workloads on single and multiple GPU nodes
Bibliographic record
Abstract
The rapid advancement of Artificial Intelligence (AI) is driving unprecedented computational demands, posing significant challenges to datacenter infrastructure and threatening the stability and resilience of modern power grids. This study presents an open-access dataset featuring a diverse set of AI training sessions recorded at sub-second resolution, designed to advance research on the energy consumption profiles of AI workloads and their interactions with power grid dynamics in datacenter environments. The dataset contains 32 training sessions on high-performance H100 and B200 8-GPU nodes and 40 sessions on consumer-grade NVIDIA GeForce RTX 3060 GPUs, encompassing over 1.8 million samples. Each session records power demand, CPU and GPU utilization, per-GPU power, memory usage, and temperature across diverse AI tasks (at the node scale, temperature refers to the GPUs temperature), including forecasting, classification, reinforcement learning, and text and image generation. Data quality was verified through detailed technical validation, including timing accuracy, hardware limit conformance, and cross-metric correlation analysis. Measurements remained within manufacturer-specified thermal and power envelopes, and observed correlations among power, utilization, temperature, and current were consistent with established processor and GPU behavior. The dataset provides a robust foundation for modeling AI datacenter energy behavior, system-level performance analysis, and power grid connection impact assessment studies.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.002 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".